What is the best approach to diagnose and treat a patient with fever of unknown origin (FUO)?

Fever of Unknown Origin: Diagnostic and Treatment Approach

Definition and Initial Classification

Fever of unknown origin (FUO) is defined as temperature >38.3°C (100.9°F) persisting for at least 3 weeks without diagnosis despite 3 outpatient visits or in-patient days, and must be classified into one of four subcategories—classical, nosocomial, neutropenic, or HIV-related—as this classification fundamentally determines your diagnostic and treatment strategy. 1, 2

The etiology distribution varies significantly by category and includes infectious, inflammatory/autoimmune, malignant, and miscellaneous causes, with geographic location also influencing the differential diagnosis. 1

Diagnostic Algorithm

Step 1: Targeted History and Physical Examination

Focus your history on these specific high-yield elements:

• Travel history: Document specific countries visited to identify region-specific diseases (malaria in Central/Western Africa, dengue and schistosomiasis in Eastern/Western Africa) 1
• Immigration status or visiting friends/relatives abroad: These patients have higher malaria rates and are less likely to seek pre-travel advice 1
• Animal exposures, occupational risks, and medication history (including recent antibiotic use) 2
• Immunosuppression history and symptoms suggesting specific organ involvement 2

Physical examination must include:

• Cardiac auscultation for new murmurs (may indicate infective endocarditis) 2
• Skin, oral cavity, conjunctival, and lymph node examination 2

Step 2: Mandatory First-Line Laboratory Testing

Order these tests before any antibiotics:

• At least 2-3 sets of blood cultures from different anatomical sites (ideally 60 mL total blood volume) 1, 2, 3
• Complete blood count with differential 1, 2
• C-reactive protein and erythrocyte sedimentation rate 2, 3
• Comprehensive metabolic panel (to identify hepatobiliary sources) 2, 3

For patients with central venous catheters, obtain simultaneous central and peripheral blood cultures to calculate differential time to positivity. 3

Step 3: First-Line Imaging

Chest radiography is the only recommended first-line imaging study, particularly if pulmonary symptoms are present or there is concern for atypical bacterial infection, HIV, tuberculosis, or malignancy. 2

Step 4: Advanced Imaging When Initial Workup is Unrevealing

[18F]FDG PET/CT is the highest-yield advanced diagnostic tool with 84-86% sensitivity and 56% diagnostic yield, and should be performed within 3 days of initiating oral glucocorticoid therapy to avoid false negatives. 1, 2, 3

Additional imaging considerations based on clinical context:

• CT chest with IV contrast identifies pulmonary sources in 72% of surgical ICU patients with suspected infection 3
• CT abdomen/pelvis with IV contrast has 81.82% positive predictive value for identifying septic foci 3
• Formal diagnostic ultrasound of abdomen for patients with abdominal symptoms or abnormal liver tests 2, 3
• CT of operative area for post-surgical patients with FUO occurring several days postoperatively 2, 3

A negative PET/CT can predict favorable prognosis through spontaneous remission and potentially allows a watchful waiting approach. 3

Step 5: Invasive Procedures When Noninvasive Tests Are Unrevealing

If noninvasive testing is unrevealing, tissue biopsy is the invasive test of choice due to relatively high diagnostic yield, which may include liver, lymph node, temporal artery, skin, skin-muscle, or bone marrow biopsy depending on clinical indications. 4

Treatment Approach

General Principles

Avoid empiric antibiotics or steroids in stable, non-neutropenic patients with FUO, as they obscure diagnosis and may be harmful if malignancy or certain infections are present; instead, rule out potentially severe causes while allowing self-limiting viral illnesses to resolve without treatment (up to 75% of cases resolve spontaneously). 1, 2, 4

Exceptions Requiring Empiric Therapy

Consider empiric therapy only in these specific situations:

• Neutropenic patients: Initiate prompt broad-spectrum antibiotics with antipseudomonal activity (high-risk patients with neutropenia <100 cells/mm³ expected >7 days should receive monotherapy with β-lactam antibiotic such as piperacillin-tazobactam as first-line) 2, 3
• Suspected tickborne rickettsial diseases 1
• Critically ill patients 1, 4

For neutropenic patients with persistent fever under antibiotics, fever alone rarely constitutes an indication to modify the antibiotic regimen, and empirical addition of vancomycin should not be done without microbiological documentation. 3

Surveillance in Treated Patients

Daily surveillance of all neutropenic patients under antibiotics is required to evaluate response, adverse effects, emergence of secondary infections, and development of resistant organisms, including daily physical examination, review of systems for new symptoms, and cultures of suspect sites. 3

Re-evaluation should occur after 2-4 days of empirical antibiotic treatment, during which most modifications of the initial regimen will be made. 3

Special Population Considerations

HIV-Positive Patients

Test CD4+ T-cell counts and HIV viral load to assess immune function, and consider opportunistic infections such as Mycobacterium tuberculosis, cytomegalovirus, and Pneumocystis jirovecii. 2

Neutropenic Patients

Appropriate imaging includes chest radiography, CT paranasal sinuses, CT chest, CT abdomen/pelvis, and FDG-PET/CT (though avoid routine sinus CT without localizing symptoms as abnormalities are common but non-discriminatory). 2, 3

Antifungal therapy should be considered only in critically ill patients with persistent fever and new pulmonary infiltrates. 3

Pediatric Patients

Medical management without imaging is usually appropriate for febrile infants >3 months and children 3-36 months without respiratory signs, but chest radiography may be appropriate if the child is ill-appearing. 2

Critical Pitfalls to Avoid

• Do not diagnose "FUO" prematurely in resource-limited settings; always assume and treat infection when fever cannot be explained by other pathologies to prevent missing treatable infections 1, 3
• Do not use high-dose steroids empirically, as they increase hospital-acquired infection risk, hyperglycemia, gastrointestinal bleeding, and delirium without improving mortality, and also mask inflammatory findings on subsequent imaging 1, 2, 3
• Avoid NSAIDs as they impair renal and coagulation function and increase stress ulcer risk 2, 3
• Obtain blood cultures before initiating any antibiotics to avoid masking the underlying cause 2, 3
• Do not systematically remove central venous catheters in clinically stable patients unless there is microbiological evidence of catheter-related infection 3
• Remember that most FUO cases result from uncommon presentations of common diseases, not rare diseases 2, 4
• Avoid non-directed antibiotic changes; do not switch from one empirical monotherapy to another or add an aminoglycoside without clinical or microbiological data justifying spectrum extension 3

Temperature Measurement Considerations

Use central temperature monitoring via pulmonary artery catheter thermistors, bladder catheters, or esophageal balloon thermistors as the preferred method for accurate diagnosis in the inpatient setting; oral or rectal temperatures are alternative reliable methods when central monitoring is unavailable. 1

Do not use axillary, tympanic membrane, temporal artery thermometers, or chemical dot thermometers for diagnostic purposes due to their unreliability. 1