I now have strong, authoritative content from Harrison's 22E (2025) and Frameworks for Internal Medicine. Let me compile the full answer.
Antibiotics for Persistent/Prolonged Fever
This is one of the most important points in clinical medicine: persistent fever alone is NOT an automatic indication for antibiotics. Here is the evidence-based framework from Harrison's Principles of Internal Medicine (22nd Ed., 2025) and other authoritative sources.
What Is "Persistent Fever"?
A fever lasting
≥3 weeks with temperature ≥38.3°C (≥101°F) on at least 2 occasions, in a non-immunocompromised patient without a clear cause after initial workup, meets the definition of
Fever of Unknown Origin (FUO) -
Harrison's 22E, p. 198. Most shorter fevers either resolve spontaneously or develop clear features that point to a cause.
Three Major Causes of Persistent Fever
Before choosing any antibiotic, a cause must be sought. The main categories are:
| Category | Median % of FUO Cases (Europe) |
|---|
| Infections | ~15-36% |
| Noninfectious inflammatory diseases (lupus, RA, vasculitis, Still's disease) | ~15-25% |
| Malignancy (lymphoma, leukemia) | ~10-15% |
| No diagnosis found | ~25-50% |
The key insight: over half of persistent fevers are NOT caused by infections - so blanket antibiotic use is often wrong and harmful.
When Antibiotics ARE Indicated
Antibiotics should be used only when there is a specific diagnosis or a specific clinical indication. Harrison's 22E explicitly states:
"Rational treatment is based on the final diagnosis. Empirical therapeutic trials should be avoided, except in cases in which a patient's condition is rapidly deteriorating."
Definite indications for empiric antibiotics in persistent fever:
- Hemodynamic instability / sepsis - broad-spectrum empiric coverage is appropriate (e.g., piperacillin-tazobactam ± vancomycin, or carbapenem for high-risk patients)
- Neutropenia (febrile neutropenia) - empiric broad-spectrum antibiotics are standard of care
- Strong suspicion of a specific infection (see below)
Antibiotics for Specific Infectious Causes of Persistent Fever
When a diagnosis is established or strongly suspected:
| Infection | First-Line Antibiotic |
|---|
| Intra-abdominal abscess (most common infectious FUO cause) | Piperacillin-tazobactam or cefazolin + metronidazole, plus drainage |
| Infective endocarditis | Based on cultures - Penicillin G/amoxicillin (Streptococci); Nafcillin/vancomycin (Staph); guided by echo + cultures |
| Tuberculosis (extrapulmonary/miliary) | HRZE (Isoniazid + Rifampicin + Pyrazinamide + Ethambutol) for 6-9 months - only after cultures/PCR collected |
| Typhoid / enteric fever | Azithromycin (uncomplicated); Ceftriaxone (severe/complicated); Ciprofloxacin (where susceptible) |
| Brucellosis | Doxycycline + Rifampicin for 6 weeks |
| Q fever (Coxiella) | Doxycycline 100 mg twice daily |
| Leptospirosis | Doxycycline or Penicillin G |
| Osteomyelitis | Nafcillin/cefazolin (MSSA); Vancomycin (MRSA); 4-6 weeks duration |
| Renal/perinephric abscess | Fluoroquinolone (e.g., ciprofloxacin) or aminoglycoside + drainage |
| Cat-scratch disease (Bartonella) | Azithromycin (5 days) |
Source: Harrison's Principles of Internal Medicine 22E; Frameworks for Internal Medicine
Empiric Antituberculous Trial
A special case: if TST/IGRA is positive, OR granulomatous disease is present with no clear alternative, OR the patient is from a TB-endemic region with a compatible picture - a trial of anti-TB therapy may be started, but only after mycobacterial cultures and PCR have been collected first. If fever does not respond after 6 weeks, reconsider the diagnosis. - Harrison's 22E, p. 198
What NOT to Do
- Do NOT start empiric broad-spectrum antibiotics in a stable patient just because the fever persists - this masks cultures, destroys diagnostic ability, creates resistance, and misses non-infectious causes (which account for >50% of FUO cases)
- Do NOT use glucocorticoids empirically until infection and lymphoma have been reasonably excluded - steroids can allow infections to spread while masking fever
The Right Approach: Workup First
Before any antibiotic:
- Full blood count, CRP, ESR, LDH, ferritin
- Blood cultures x3 (before any antibiotic)
- Urine analysis + culture
- Chest X-ray
- CT abdomen/pelvis (high-yield - detects abscesses, lymphadenopathy, masses)
- ANA, RF, HIV, CMV, EBV, TST/IGRA based on history
- Echocardiogram if endocarditis suspected
Only after this workup points toward a specific infectious cause should antibiotics be chosen.
Bottom line: Persistent fever requires a diagnosis, not just an antibiotic. Antibiotics are indicated when you find an infectious cause, when the patient is septic/hemodynamically unstable, or when neutropenia is present. For stable patients with unexplained FUO, the majority will either become symptom-free spontaneously or reveal a non-infectious diagnosis on thorough workup.