A Multicenter Randomized Trial of Continuous versus Intermittent β-Lactam Infusion in Severe Sepsis

Dulhunty. Am J Respir Crit Care Med. First published online 22 Jul 2015. DOI: 10.1164/rccm.201505-0857OC

Clinical Question

• In adult ICU patients on beta-lactam antibiotics, does continuous infusion compared to intermittent bolus increase survival?

Design

• Randomised, controlled trial
• Double-blind with double-dummy preparations
  • Unblinded staff prepared drugs and consecutively labelled sealed opaque envelopes
  • Study participants, clinicians and data investigators remained blinded
• Pilot trial conducted to guide power calculation
  • Sample size of 420 required to achieve 90% power to detect a 3 day difference in the primary outcome
• Intention-to-treat (ITT) analysis for primary end points, in addition to:
  • Modified intention-to-treat (mITT) conducted in eligible patients who received a study drug
  • A priori per protocol analysis in those who received 3 or more days of a study drug

Setting

• 25 Intensive Care Units (17 in Australia, 7 in New Zealand and 1 in Hong Kong)
• July 2012 to April 2014

Population

• Inclusion: adults with severe sepsis commenced on any of:
  • Piperacillin-tazobactam
  • Ticarcillin-clavulanate
  • Meropenem
• Exclusion:
  • Received beta-lactam for more than 24 hours duration prior to randomisation
  • Pregnancy
  • Allergy to study drugs
  • No central venous line with >= 3 lumens
  • Palliative treatment only, limits to care, death is imminent or death is likely within 90 days due to underlying process
• 2630 patients screened, 432 eligible patients were randomised (ITT), 422 received a study drug (mITT)
  • Pathogen isolated in blood cultures in 83 patients

Intervention

• Continuous infusion of beta-lactam antibiotics
  • Loading dose of chosen antibiotic
  • Continuous infusion thereafter
    • 1g in 100ml Meropenem over 24 hours
    • 13.5g in 250ml Piperacillin-tazobactam over 24 hours
    • 12.4g in 250ml Ticarcillin-clavulanate over 24 hours
  • Administered with a normal saline dummy intermittent infusion

Control

• Intermittent infusion of beta-lactam antibiotics
  • Loading dose of chosen antibiotic
  • Intermittent infusion thereafter
    • 1g in 20ml Meropenem once daily
    • 4.5g in 20ml Piperacillin-tazobactam three times daily
    • 3.1g in 20ml Ticarcillin-clavulanate four times daily
  • Administered with a normal saline dummy continuous infusion

Management common to both groups

• Infusions were administered via a primed central venous line
• A change between the three beta-lactam antibiotics or to flucloxacillin was permitted within 14 days
  • If flucloxacillin was introduced, a non-blinded loading dose was administered followed by blinded administration as per the patient’s allocation
• Total dose over 24 hours was the same regardless of group allocation

Outcome

• Characteristics of the groups (continuous vs intermittent)
  • APACHE II score: 21 [IQR 17-26] vs 20 [IQR 16-25]
  • Median antibiotic administration prior to randomisation: 13 hours vs 12 hours
  • Median duration of blinded study drug treatment: 3.2 days [IQR 1.9-6.0] vs 3.7 days [IQR 1.9-5.9]
  • Median total course of beta-lactam antibiotic: 5.3 days [IQR 2.9-7.7] vs 5.0 days [IQR 3.1-8.0]
  • Change in antibiotic: 9.4% patients vs 11.8% patients
  • Choice of antibiotic:
    • Piperacillin-tazobactam: 69.3% vs 71.4%
    • Meropenem: 29.7% vs 27.3%
    • Ticarcillin-clavulanate: 0.9% vs 1.4%
• Primary outcome: no difference was demonstrated in alive ICU-free days at 28 days
  • Continuous group: 18 days [IQR 2-24]
  • Intermittent: 20 days [IQR 3-24]
  • P-value: 0.38
• Secondary outcome:
  • ICU-free days at 90 days: no difference found
    • Hazard ratio: 0.91 (95% CI 0.63-1.31; p-value = 0.61)
  • Clinical cure at 14 days after antibiotic cessation: no difference found
    • Odds ratio: 1.12 (95% CI 0.77-1.63; p-value 0.56)
• Modified Intention-to-treat and per-protocol analysis
  • Differences remained non-significant for these two a priori defined populations

Authors’ Conclusions

• In a heterogenous Intensive Care Unit population, continuous and intermittent infusion of beta-lactams appear to have equivalent outcomes

Strengths

• Inclusion of a pragmatic, heterogenous ICU population including those on renal replacement therapy
• Appropriate randomisation, concealment and blinding
• Double-blind and double-dummy administration removes possible bias, which is particularly important given that the treating clinicians could influence the outcome measure (early or late discharge from ICU)

Weaknesses

• Not powered to detect mortality differences; this study used alive ICU-free days at 28 days as a surrogate
• Relatively short antibiotic courses were used, and so may not represent all ICU practice
• Very few patients received ticarcillin-clavulanate, so the results may not be representative for this antibiotic
• Clinical cure was assessed subjectively by blinded clinician

The Bottom Line

• In ICU patients, with underlying pathophysiological changes of sepsis and organ dysfunction, there appears to be no benefit from continuous infusion of beta-lactam antibiotics compared to intermittent bolus infusion
• Given the simplicity of an intermittent bolus infusion, this should be the preferred method for these antibiotics in this patient group

External Links

• [article] A multicenter randomized trial of continuous versus intermittent beta-lactam infusions in severe sepsis
• [further reading] Alternative dosing strategies for intravenous antibiotics to treat severe infections
• [further reading] LITFL CCC – Antimicrobial dosing and kill characteristics

Metadata

Summary author: @DuncanChambler
Summary date: 2 October 2015
Peer-review editor: @avkwong