Use of procalcitonin to reduce patients’ exposure to antibiotics in Intensive Care Units

Bóuadma. Lancet 2010; 375:463-74. doi:10.1016/S0140-6736(09)61879-1

Clinical Question

• In adult patients with suspected bacterial infections, does procalcitonin-guided antibiotic care compared to usual care reduce patient exposure to antibiotics without increasing mortality?

Design

• Prospective, parallel group, open-label, randomised, controlled trial
• Centralised, computer-generated randomisation sequence with adequate masking
• Stratified by centre in blocks of 2, 4 or 6
• Investigators were unaware of aggregate outcome during trial period but were not blinded to individual patient allocation (i.e. open-label)
• Primary endpoint objective and not patient reported to limit bias from open-label design
• Classification of infective episodes included clinical, radiographic and microbiological information, and were adjudicated by 4 expert clinicians who remained blinded to allocation
• Intention-to-treat analysis with worst-case imputation for missing data
• Designed as superiority trial in terms of antibiotic use
  • With 266 patients, it was powered to detect a 3 day reduction in antibiotic use, assuming mean of 12 days in control group, with 90% power and two-sided α=0.05 testing
• Also designed as a non-inferiority trial in terms of death
  • With 300 patients, it was powered to exclude a 10% non-inferiority margin, assuming a 35% mortality in control group, with 80% power and α=0.10

Setting

• 8 intensive care units across France
  • 5 medical units vs 2 surgical units vs 1 mixed unit
  • Within 5 teaching hospitals and 1 general hospital
• June 2007 to May 2008

Population

• Inclusion: Adults with suspected bacterial infections at admission or during their stay in intensive care who had:
  • Not received antibiotics before inclusion
  • Or had received antibiotics for less than 24 hours but were within 12 hours of admission
• Exclusion:
  • Under 18 years of age
  • Pregnancy
  • Anticipated less than 3 days admission to intensive care
  • Bone-marrow transplant or chemotherapy
  • Indications for long-term antibiotics (e.g. infective endocarditis)
  • Poor chance of survival (i.e. SAPS 2 > 65 points) or existing do-not-resuscitate order
• 1315 patients assessed, 630 enrolled, 621 included in analysis (1.4% lost / withdrawn)

Intervention

• Procalcitonin-based protocolised antibiotic care
  1. Starting antibiotics
     • According to baseline serum procalcitonin level
       • < 0.25 µg/l → antibiotics strongly discouraged
       • ≥ 0.25 and < 0.5 µg/l → antibiotics discouraged
       • ≥ 0.5 and < 1.0 µg/l → antibiotics encouraged
       • ≥ 1.0 µg/l → antibiotics strongly encouraged
     • Repeated 6–12 hours later if antibiotics not started to detect late peak
  2. Stopping antibiotics
     • According to serial serum procalcitonin level protocol
     • < 0.25 µg/l → stopping antibiotics strongly encouraged
     • ≥ 0.25 and < 0.5 µg/l or ≥ 80% decrease from peak → stopping antibiotics encouraged
     • ≥ 0.5 µg/l and < 80% decrease from peak → continuing antibiotics encouraged
     • ≥ 0.5 µg/l and increase from peak→ changing antibiotics strongly encouraged
  • Serum sampling was performed:
    • At inclusion
    • At each infectious episode until day 28
    • At each morning if receiving antibiotics
    • Using Brahms’ Krypton procalcitonin assay (minimum detection 0.06 µg/l with imprecision of 6–10%)
    • And reported back to clinicians within 2 hours

Control

• Clinician decision antibiotic care
  • International and local antibiotic guidelines provided as a reminder and gold-standard
  • Clinicians decided optimum antibiotic management including duration

Management common to both groups

• Drug selection was at the discretion of clinicians
• Clinicians were encouraged to start antibiotics as soon as possible in severe infections
• De-escalation from broad-spectrum to narrow-spectrum antibiotics as soon as relevant information was available

Outcome

• Primary outcome:
  • Superiority
    • The absolute difference in antibiotic exposure was 2.7 days (23% less) favouring the procalcitonin group
    • 95% CI 1.4 to 4.1 days
    • p-value <0.0001
  • Non-inferiority
    • The absolute difference in mortality at 28-days was 0.8% and at 60-days was 3.8% favouring the control group
    • 90% CI limits did not cross the pre-specified 10% non-inferiority margin
      • Death at 28-days: 90% CI -4.6 to 6.2%
      • Death at 60-days: 90% CI -2.1 to 9.7%
    • Hazard ratio of death by 60-days was 0.96 (90% CI 0.84–1.09)
• Secondary outcome:
  • No statistically significant differences were identified for most of the secondary outcomes (see table for details)
    • Relapse or superinfection
    • Days without mechanical ventilation
    • Length of stay in ICU
    • Length of stay in hospital
    • Emergence of multi drug-resistant bacteria
  • Statistically significant differences were identified for the following, favouring the procalcitonin group
    • Days of exposure to each antibiotic per 1000 inpatient days
    • Duration of antibiotic treatment according to infection site

Authors’ Conclusions

• Procalcitonin-guided antibiotic therapy reduced antibiotic exposure by 23% and is non-inferior to standard care regarding mortality outcomes

Strengths

• Design – the combination of superiority and non-inferiority allowed two important outcomes to be investigated appropriately
• Internal validity – the appropriate intention-to-treat analysis, minimal loss / withdrawal and worst-case imputation all increase the validity of the results and reduce the chance of a false positive outcome (at the cost of increased chance of false negative outcome)
• External validity – the multi-centre design and inclusion of many pathologies increases the generalisability to wider populations

Weaknesses

• Non-inferiority design – a 10% non-inferiority margin is in line with existing antibiotic trials and guidance, but it could be debated as too extreme to conclude ‘clinical non-inferiority’.
• Internal validity – in a high proportion of cases clinicians overruled the procalcitonin protocol suggesting it provides useful advice but not rules for antibiotic use; consequently, outside trials the benefits of procalcitonin may not be as strong
• External validity – only 10% were surgical patients, reducing validity in this population; neutropenic patients and those receiving chemotherapy / bone marrow transplants were excluded, so the results cannot be extrapolated to these populations

The Bottom Line

• This trial suggests that procalcitonin-guided antibiotic therapy can reduce antibiotic administration without a significant effect on mortality outcomes

External Links

• [article] Use of procalcitonin to reduce patients’ exposure to antibiotics in intensive care units (PRORATA trial): a multicentre randomised controlled trial
• [further reading] Procalcitonin testing to guide antibiotic therapy for the treatment of sepsis in intensive care settings and for suspected bacterial infection in emergency department settings: a systematic review and cost-effectiveness analysis
• [further reading] Procalcitonin by LITFL

Metadata

Summary author: @DuncanChambler
Summary date: 4 December 2015
Peer-review editor: @avkwong