01. Fever of Unknown Origin

Resident Editor:               John Landefeld, M.D.

Faculty Editor:                    Harry Hollander, M.D.

BOTTOM LINE

✔ The key to diagnosis is careful and repeated H&P.

✔Infection, neoplasia, and non-infectious inflammatory disease are the most common causes of classic FUO.

✔ No guidelines exist, but minimum obligatory investigations should include imaging and non-infectious serologic tests if initial evaluation is nondiagnostic.

✔ Empiric therapy is warranted only for unstable or immunocompromised patients, or if there is strong suspicion of potentially rapidly progressive disease (e.g., TB, SBE, GCA).


Background/Epidemiology

  • Classic: temperature >38.3 C (100.9 °F), duration >3 weeks, ≥3 outpatient visits or ≥3 days of inpatient investigation has not revealed source.
  • Immune deficient (neutropenic): temperature >38.3 C, neutrophil count of ≤500 per mm3, ≥3 day evaluation has not revealed source.
  • Nosocomial: temperature >38.3 C, patient hospitalized ≥24 hours but no fever on admission and no obvious source of fever prior to admission, ≥3 day evaluation has not revealed source.
  • HIV-associated: temperature >38.3 C, duration >4 weeks for outpatient or ≥3 days for inpatient, confirmed human immunodeficiency virus infection.
  • Prevalence of classic fever of unknown origin in hospitalized patients is 1-3%.
  • Incidence is declining due to improved diagnostics.
  • Diagnoses are more likely to be made in patients >65 years.
  • There are over 200 causes of fever of unknown origin.
  • Most often an atypical presentation of a common disease.

Signs and Symptoms

  • Often vague and/or intermittent.
  • Pathology of any organ system can cause fever of unknown origin, therefore use signs and symptoms among other clues to guide diagnostic approach.
  • The longer the duration, the less likely infection is an etiology.
  • May be helpful for distinguishing etiology, for example:
    • Rigors are virtually never a component of rheumatic/inflammatory fevers

Differential Diagnosis

A. Causes by fever type

  • Classic: infection, neoplasia, non-infectious inflammatory disease (see below).
  • Immune deficient: opportunistic bacterial infections, aspergillosis, candidiasis, herpes simplex.
  • Nosocomial: drug fever, Clostridium difficile, septic thrombophlebitis, pulmonary embolism, sinusitis.
  • HIV-associated: >75% infectious, including bacterial infections, pneumocystis pneumonia, cytomegalovirus, Mycobacterium avium, and other infections; drug fever, Kaposi’s sarcoma, lymphoma. Rarely due to human immunodeficiency virus alone.

B. Most common causes of classic FUO

  • Infection (15-30%): tuberculosis (esp. extrapulmonary/disseminated), abscess (abdominal, pelvic or dental), endocarditis (consider “HACEK” organisms, Bartonella, Legionella, Coxiella), osteoarticular infections, typhoid/enteric fever, endemic fungi, brucellosis, leishmaniasis, prostatitis, malaria, rickettsial infection, chronic sinusitis, human immunodeficiency virus, Lyme disease, cat-scratch disease, Epstein-Barr virus, cytomegalovirus.
  • Neoplasia (10-30%): lymphoma, leukemia, myelodysplastic syndromes, hepatoma, hepatic metastases, renal cell carcinoma.
  • Non-infectious inflammatory disease (33-40%): Adult Onset Still’s Disease, autoimmune hepatitis, systemic vasculitis, temporal arteritis/polymyalgia rheumatica (age >65), rheumatoid arthritis, inflammatory bowel disease, reactive arthritis, sarcoidosis, myositis, mixed cryoglobulinemia.
  • Miscellaneous (5-14%): drug fever, factitious, deep vein thrombosis/pulmonary embolism, hepatitis (alcoholic, granulomatous, lupoid), hyperthyroidism/subacute thyroiditis, adrenal insufficiency, periodic fever syndromes, hematoma.
  • Undiagnosed FUO (20-30%): no diagnosis made despite extensive workup, most will resolve spontaneously within weeks to months with good prognosis even if no clearcut diagnosis made (5-year mortality 3.2%).
     

C. Drug fever – can occur at any time during drug therapy, patient may be unaware of fevers, usually resolves 2-3 days after discontinuing culprit med though may take up to 1 week, temp-pulse disparity (i.e. relative bradycardia), eosinophilia and rash present in only 25%.

  • Agents commonly associated with drug-induced fever include but are not limited to: allopurinol, captopril, cimetidine, clofribrate, erythromycin, heparin, hydralazine, hydrochlorothiazide, isoniazid, meperidine, methyldopa, nifedipine, nitrofurantoin, penicillin, phenytoin, procainamide and quinidine

Evaluation

A. History

  • Duration (prolonged fever is less likely to be infectious), travel history, country of origin, sick contacts, other exposures (i.e. tick/pet/farm animal exposure, occupational, unpasteurized foods, rare meats), meds, prior surgeries, cancer history; history of tuberculosis, rheumatic fever, or endocarditis; prior transfusions, indwelling foreign body (i.e. pacemaker, prosthetic joint, cosmetic implant), psych history (factitious), intravenous drug use, sexual history, full ob/gyn history in women.

B. Physical Exam

  • Comprehensive physical exam, often needs to be repeated on separate visits.
    • Infectious etiologies:
      • Morning temperature spikes seen in miliary TB, typhoid/enteric fever, Whipple’s disease
      • Faget sign (relative bradycardia with fever) seen in typhoid, malaria, babesiosis, ehrlichiosis, Q fever
      • Double quotidian fevers can be seen in miliary TB, malaria, or visceral leishmaniasis
      • Isolated hepatomegaly suggests Q fever, typhoid fever, visceral leishmaniasis, brucellosis, rat-bite fever, or relapsing fever
    • Neoplastic etiologies:
      • Significant weight loss (>2 lbs/week), post-hot-bath pruritus
    • Rheumatologic/Inflammatory etiologies:
      • Chills/rigors virtually never occur with rheumatologic/inflammatory fevers
      • Dry cough can be clue to temporal arteritis
      • Oral ulcers suggest Behcet’s, wherease diffuse lymphadenopathy suggests Still’s
      • Acalculous cholecystitis may occur in FUO due to polyarteritis nodosa

C. Labs and Tests

  • No established guidelines or gold standard.
  • If no diagnostic clues, start with CBC w/ diff, smear, lytes, LFTs, blood cultures x3 (hold for slow-growing organisms), UA and culture, ESR/CRP, PPD, HIV, CK, CXR; can also consider abominal/pelvic ultrasound or CT, RF, ANA, ANCA, and LDH at this stage of workup. LP is often recommended but of questionable value in absence of neurologic/meningeal findings.
  • Lymph node biopsies can be pursued, however inguinal node biopsies should be avoided because they are often nonspecific
  • If still no clear source, and above studies do not provide clues, then additional investigations should be considered on a case-by-case basis depending upon epidemiology and clinical findings.
  • Increasing evidence for utility of CT/FDG-PET earlier on in the investigation, with sensitivity of 56-100%, specificity of 75-81%, negative predictive value of >99%.

Treatment

  • Discontinue unnecessary medications.
  • If no source identified after extensive investigation, then watchful waiting is reasonable for patients who are clinically well, with 5-year mortality in undiagnosed fever of unknown origin of 3.2%.
  • Empiric therapy with antibiotics or corticosteroids is discouraged, unless patient is clinically declining, immunocompromised, or a potentially rapidly progressive diagnosis is strongly suspected (e.g., subacute bacterial endocarditis with peripheral manifestations, miliary tuberculosis awaiting biopsy results, giant cell arteritis).
    • If pursuing empiric therapy, strongly consider specialty referral (i.e. ophthalmology referral for temporal artery biopsy if concerned for giant cell arteritis) and make sure to obtain blood cultures x3 (with hold for slow growing organisms) before starting antibiotic therapy
  • Consider antipyretics if the patient has associated symptoms (headache, arthralgia, myalgia), although they may alter fever patterns.
  • A well defined short course of empiric therapy may have diagnostic utility when clinical suspicion is high (i.e. low dose steroids as empiric therapy for polymyalgia rheumatica)

When to Refer

  • If no clear etiology is found after extensive workup, consider referral to infectious disease, rheumatology, or hematology/oncology.

References

Cunha BA. Fever of unknown origin: focused diagnostic approach based on clinical clues from the history, physical examination, and laboratory tests. Infect Dis Clin North Am 2007;21(4):1137-87, xi.

Cunha BA, Lortholary O, Cunha CB. Fever of unknown origin: A Clinical Approach. Am J Med 2015;128(10):1138e.1-1138e.15.

Hayakawa K, Ramasamy B, Chandrasekar PH. Fever of unknown origin: an evidence-based  review. Am J Med Sci. 2012; 344(4):307-16.

Mourad O, Palda V, Detsky AS. A comprehensive evidence-based approach to fever of unknown origin. Arch Intern Med 2003;163(5):545-51.

Roth AR, Basello GM. Approach to the adult patient with fever of unknown origin. Am Fam Physician 2003;68(11):2223-8.

Tolia J, Smith LG. Fever of unknown origin: historical and physical clues to making the diagnosis.  Infect Dis Clin North Am 2007;21(4):917-36, viii.

Vanderschueren S, Knockaert DC, Peetermans WE, Bobbaers HJ. Lack of value of the naproxen test in the differential diagnosis of prolonged febrile illnesses. Am J Med. 2003;115(7):572-5.

Varghese GM, Trowbridge P, Doherty T. Investigating and managing pyrexia of unknown origin in adults. BMJ. 2010;341:C5470.