How should C‑reactive protein (CRP) be utilized in the assessment and management of an adult or adolescent presenting with an acute fever of unknown origin?

Use of CRP in Acute Febrile Illness

In patients with acute fever of unknown origin, measure CRP in conjunction with clinical assessment when the probability of bacterial infection is low to intermediate, but do not rely on CRP alone to rule out bacterial infection when clinical suspicion is high. 1

Clinical Decision Algorithm

Step 1: Assess Pre-test Probability of Bacterial Infection

Low to Intermediate Probability:

• Measure CRP alongside bedside clinical evaluation 1
• CRP can help guide antibiotic stewardship decisions
• Consider obtaining baseline value to assist with later antibiotic discontinuation 1

High Probability:

• Do NOT measure CRP to rule out bacterial infection 1
• Initiate appropriate empiric antibiotics immediately
• CRP should not delay treatment decisions

Step 2: Interpret CRP Values with Critical Context

Timing Matters Critically:

• CRP rises 12-24 hours after inflammatory/infectious insult 1
• Reaches maximum at 48 hours 1
• CRP obtained <12 hours from fever onset is unreliable for distinguishing bacteremia 2
• Diagnostic accuracy improves significantly when measured >24 hours after fever onset 3

Standard Cutoff Values:

• Normal: <5 mg/L 1
• Typical diagnostic threshold: 10 mg/L 1
• Adjust cutoffs based on fever duration:

  • 12-24 hours: 6 mg/dL 3

  • 24-48 hours: 10.7 mg/dL 3

  • 48 hours: 12.6 mg/dL 3

Step 3: Understand CRP Limitations

CRP Has Modest Diagnostic Accuracy:

• Area under ROC curve of only 0.73 for sepsis diagnosis 1
• Sensitivity 80% but specificity only 61% 1
• Procalcitonin (PCT) demonstrates superior diagnostic accuracy (AUC 0.85 vs 0.73) 1

Confounding Factors That Elevate CRP:

• Obesity and lifestyle factors 4
• Fatty liver disease 4
• Smoking 4
• Non-infectious inflammatory conditions (often higher than infections) 5, 6, 7
• Rheumatological diseases 6, 7

Factors That Blunt CRP Response:

• Neutropenia 1
• Immunodeficiency 1
• NSAIDs 1

Etiology-Specific Patterns

When CRP is Elevated in FUO:

• Higher CRP correlates with increased odds of diagnosis (OR 2.11 per log increase) 8
• Paradoxically, CRP may be HIGHER in non-infectious causes (autoimmune/inflammatory diseases) than infections 5, 7
• Very high CRP (>150 mg/L) does NOT reliably predict bacteremia 2

Diagnostic Clues by Disease Category:

• Infections: Moderate CRP elevation, positive T-SPOT.TB for tuberculosis 7
• Autoimmune/Inflammatory: Markedly elevated CRP, ESR, ferritin, IL-6; low albumin 6, 7
• Malignancy: Variable CRP; ESR >40 mm/h in ALL cases 9; elevated LDH and ferritin 7

Critical Pitfalls to Avoid

1. Do not use CRP alone to exclude bacterial infection - it has insufficient negative predictive value 1

2. Do not interpret early CRP values - wait at least 12-24 hours from fever onset for meaningful results 2, 3

3. Do not assume high CRP = bacterial infection - non-infectious inflammatory diseases often produce higher CRP than infections 5, 7

4. Do not withhold antibiotics in sepsis based on biomarkers - clinical judgment supersedes CRP values when bacterial infection probability is high 1

5. Recognize CRP >10 mg/L is not always acute infection - chronic conditions (obesity, smoking, inflammatory diseases) can produce sustained elevation 4

Practical Application

For Antibiotic Stewardship:

• Use CRP trends (not single values) to guide de-escalation in stabilized patients 1
• Consider PCT over CRP for antibiotic discontinuation decisions (stronger evidence base) 1

For Diagnostic Workup:

• Combine CRP with complete blood count, ESR, ferritin, and LDH for pattern recognition 9, 7
• Normal ESR appears to exclude malignancy powerfully 9
• Consider PET/CT when CRP elevated but diagnosis remains elusive 8

In Nosocomial Fever:

• CRP has particularly limited utility in critically ill ICU patients 5
• Do not rely on CRP to distinguish infectious from non-infectious nosocomial fever 5