In a newborn receiving empiric antibiotics for suspected sepsis, at what C‑reactive protein (CRP) level or trend should the antibiotic therapy be prolonged?

CRP-Guided Antibiotic Duration in Neonatal Sepsis

Discontinue antibiotics when CRP falls below 10 mg/L at 24–48 hours after initiation, provided blood cultures remain negative and the infant shows clinical improvement. This approach safely identifies 99% of neonates who do not require further antibiotic therapy and reduces unnecessary antibiotic exposure from a median of 5 days to 3–4 days 1, 2.

The Evidence-Based CRP Threshold

• A CRP level <10 mg/L measured 24–48 hours after starting antibiotics has a 99% negative predictive value for ruling out bacterial infection, meaning antibiotics can be safely stopped at this point in culture-negative cases 1, 2.

• If the initial CRP at 24–48 hours is ≥10 mg/L, continue antibiotics and repeat CRP daily until it falls below 10 mg/L, at which point therapy can be discontinued 1.

• Serial CRP measurements at baseline, 24,36, and 48 hours predict antibiotic sensitivity: a declining CRP by 48 hours indicates the organism is sensitive to empirical therapy (89% sensitivity, 80% specificity), while rising or persistently elevated CRP suggests resistance and warrants escalation 3.

Integration with Guideline-Based Reassessment

The Surviving Sepsis Campaign guidelines mandate daily clinical and laboratory assessment for de-escalation after 48 hours, with review of ongoing indication for empirical therapy guided by microbiologic results and clinical improvement 4. CRP provides the objective laboratory parameter to operationalize this recommendation, replacing subjective clinical assessments that lead to unnecessary prolongation 5.

Specific Algorithm for Decision-Making:

At 24–48 hours after antibiotic initiation:

• CRP <10 mg/L + negative blood culture + clinical improvement → Stop antibiotics 1, 2
• CRP ≥10 mg/L → Continue antibiotics, repeat CRP daily 1
• CRP declining by 48 hours → Current antibiotics are appropriate, continue until CRP <10 mg/L 3
• CRP rising or persistently elevated at 48 hours → Escalate antibiotic coverage immediately (consider amikacin plus cloxacillin or vancomycin plus ceftazidime for nosocomial pathogens) 6, 7, 3

At 48–72 hours:

• Cultures negative + CRP <10 mg/L + clinical improvement → Discontinue antibiotics to avoid unnecessary exposure 4, 8, 6
• Cultures positive → De-escalate to narrowest effective spectrum based on susceptibilities, continue until CRP normalizes 4

Clinical Context and Common Pitfalls

The most common mistake is continuing antibiotics beyond 48 hours based on subjective clinical assessments (pale appearance, behavioral change) or arbitrary CRP interpretations when cultures are negative 5. A 2024 study found that 36% of neonates with negative cultures received prolonged therapy, with clinicians citing "clinical appearance" (38%) and "elevated CRP" (36%) as reasons—yet these were not supported by guideline recommendations 5.

Avoid these specific errors:

• Do not continue antibiotics for isolated respiratory symptoms (tachypnea, hypoxia) if CRP is <10 mg/L at 24–48 hours and cultures are negative 5.
• Do not interpret a single elevated CRP at baseline as justification for prolonged therapy—the trend matters more than the absolute value 1, 3.
• Do not wait for a full 5-day course when CRP normalizes earlier—this increases antibiotic exposure without reducing relapse rates (mean 3.7 days CRP-guided vs. 5.5 days fixed duration, with equivalent relapse rates of <1%) 1.

Special Populations Requiring Caution

Premature infants <33 weeks gestation may require closer monitoring, as one study reported a single false-negative case in a 31-week infant and several re-evaluations in infants ≤33 weeks 2. However, the overall negative predictive value remained 99%, supporting the approach even in this population 2.

For nosocomial/late-onset sepsis (>72 hours of life), the same CRP thresholds apply, but empirical coverage must shift to amikacin plus cloxacillin (or vancomycin if MRSA suspected) rather than ampicillin plus gentamicin 6, 7.

Practical Implementation

**Introduce an automatic stop order at 48 hours for low-risk infants with CRP <10 mg/L and negative cultures**—this intervention increased the proportion receiving ≤48 hours of antibiotics from 19% to 73% and reduced avoidable doses (>48 hours, <5 days) from 50% to 0.8% 9.

Measure CRP at a single time point (36 hours) if resources are limited—this simplified approach still achieves significant reductions in antibiotic duration without compromising safety 9.