Skip to main content

Foundational Trials

Kress 2000

Image: Shutterstock / ame435

Paper

Sections

Publication  |  Context & Rationale  |  Design & Methods  |  Key Results  |  Internal Validity  |  External Validity  |  Strengths & Limitations  |  Interpretation & Why It Matters  |  Controversies & Subsequent Evidence  |  Summary  |  Further Reading  |  Overall Takeaway

Publication

  • Title
    • Daily interruption of sedative infusions in critically ill patients undergoing mechanical ventilation
  • Acronym
    • None
  • Year
    • 2000
  • Citation
    • Kress JP, Pohlman AS, O’Connor MF, Hall JB et al. Daily Interruption of Sedative Infusions in Critically Ill Patients Undergoing Mechanical Ventilation. N Engl J Med. 2000;342(20):1471–1477

Context & Rationale

  • Background
    • Prior to this trial, continuous intravenous sedation was standard for ICU patients receiving invasive mechanical ventilation.
    • Although sedation can improve patient comfort, prolonged or excessive sedation was suspected to increase the risks of prolonged mechanical ventilation, ICU and hospital lengths of stay, and potentially other complications.
    • Non-randomised data suggested that lighter sedation strategies might improve outcomes, but robust, controlled evidence was limited.
  • Why This Matters
    • Prolonged mechanical ventilation increases the risk of ventilator-associated complications, such as pneumonia, as well as delirium, and critical illness polyneuromyopathy.
    • If daily sedation interruption could reduce ventilator days without increasing adverse events, this would represent a significant advance in ICU practice.
    • Even a two‑day reduction in ventilation could reduce the incidence of ventilator‑associated pneumonia, delirium and excess ICU bed‑days, while also cutting the ~US $1 billion annual spend on sedatives reported for US ICUs in 1997.
  • Question
    • Would routine daily suspension of all sedative and opioid infusions shorten duration of ventilation and ICU length of stay versus usual care, where sedation breaks were optional?

Design & Methodology

Trial Design

  • Design

    • Single‑centre, concealed‑allocation, open‑label RCT with a 2 × 2 factorial structure (midazolam‑morphine vs propofol‑morphine).

  • Setting
    • A 10‑bed medical ICU, in Chicago, USA.

Population

  • Inclusion Criteria
    • Adults (>18 years old)
    • Requiring mechanical ventilation with continuous IV sedation
    • Expected to need ventilatory support for at least 48 hours
  • Exclusion Criteria
    • Severe neurologic injury requiring fixed sedation levels
    • Comfort care/withdrawal of life-sustaining treatment
    • Other major contraindications to sedation interruption (e.g., continuous neuromuscular blockade), pregnancy, post‑cardiac‑arrest admission, transfer from another hospital already on sedation.

Intervention

  • Stopping Sedation
    • From study day 3 onward, a research nurse stopped both sedative and morphine infusions each morning until patients obeyed ≥3 of 4 commands or became agitated.
  • Restarting Sedation
    • Infusions were then restarted at 50 % of the previous dose and titrated to a Ramsay Sedation Scale score of 3‑4.

Control

  • Standard Practice
    • Standard sedation practices (continuous IV sedation without routine interruption), following usual ICU protocols at the time (dose titrated for comfort, but no formal daily break).
  • Sedation Titration
    • Continuous infusions titrated by bedside staff; breaks permitted but not mandated.

Statistical Plan

  • Power Calculation

    • No formal power calculation was published.

  • Post Randomisation Exclusions
    • Patients who died or were extubated within the first two days of in the ICU were excluded from analysis
  • Analysis
    • Intention‑to‑treat; non‑parametric tests plus Cox models adjusted for baseline covariates.

Other

  • Blinding
    • Unblinded: treating clinicians and patients were aware of whether sedation was interrupted.
    • The nature of the intervention made blinding impractical.
    • Potential for performance bias, though investigators attempted to standardize care.
  • Follow Up
    • To hospital discharge or death.

Key Results

Daily Awakening (DA) vs Continuous Sedation (CS)

  • Early Stopping

    The trial was not stopped early (although there was no formal power calculation)

Primary Outcome - Duration of Mechanical Ventilatio

  • Duration of Mechanical Ventilation
    (days; median/IQR)
    • 4.9 (2.5 to 8.6) vs 7.3 (3.4 to 16.1
    • Absolute Difference: −2.4
    • HR extubation, 1.9; 95% CI, 1.3  to 2.7
    • P < 0.001

    Secondary Outcomes

    • ICU Length of Stay
      (days; median/IQR)
      • 6.4 (3.9 to 12.0) vs 9.9 (4.7 to 17.9)
      • Absolute Difference, −3.5 
      • HR discharge, 1.6; 95% CI, 1.1 to 2.3
      • P = 0.02
    • Hospital Length of Stay
      (days; median/IQR)
      • 13.3 (7.3 to 20.0) vs 16.9 (8.5 to 26.6)
      • Absolute Difference, –3.6
      • P=0.19

    • In-Hospital Mortality
      • 36 % vs 46.7 %
      • Relative Risk, 0.77
      • P=0.25
    • Self Extubation
      • 4.4 % vs 6.7 %
      • P=0.88

    Notes

    • Awakening halved benzodiazepine and opioid exposure and yielded ≈ 2.4 ventilator‑free days.
    • Safety profile comparable, with <5 % accidental device removal.
    • Mortality signal favoured intervention but confidence intervals wide.

    Internal Validity

    • Randomisation & Allocation

      Computer‑generated, sealed envelopes – adequate concealment.

    • Performance/Detection Bias
      • An unblinded study design introduces the possibility that clinicians might modify other aspects of care (performance bias).
      • Objective outcomes like ventilator days are less susceptible to detection bias.
    • Protocol Adherence
      • A daily sedation break was delivered on 100 % of eligible days
      • 30% (18/60) of control patients received unscheduled breaks, leading to a potential dilution of the effect of the intervention
    • Outcome Assessment
      • Primary outcome (duration of mechanical ventilation) was objective and clearly defined.
      • Secondary outcomes (ICU LOS, sedation drug use) also objective.
    • Statistical Rigor
      • Pre‑specified analyses were completed
      • No formal power calculation
      • No test for factorial interaction
    • Separation of the Variable of Interest

      Midazolam subgroup (n=37 & 29)

      • Midazolam: 229.8 mg vs 425.5 mg; (46% reduction; P=0.05)
      • Morphine: 205 mg vs 481 mg (57% reduction; P=0.009)

      Propofol subgroup (n=31 & 31)

      • Propofol: 15,150 mg vs 17,588 mg (P=0.54)
      • Morphine: 352 mg vs 382 mg (P=0.33)

    • Key Delivery Aspects
      • Implementation hinged on nurse and physician comfort with sedation breaks.
      • The protocol required vigilance to ensure patient safety and comfort.
    • Drop Outs or Exclusions
      • 22/150 (15 %) died or were extubated within 48 hours and were omitted from analysis.
      • Similar proportions, but risk of attrition bias.
    • Baseline Characteristics
      • Age 57 vs 61 years
      • APACHE II 20 vs 22
      • Diagnoses balanced
    • Heterogeneity

      Single-center design and homogeneous population (medical ICU) may reduce heterogeneity, although it also limits broad external validity.

    • Timing
      • Sedation breaks were conducted daily.
      • Timing was standardized to mornings.
    • Dose

      Intervention:

      • Full cessation of sedation daily.
      • Re-initiation at half dose if the patient became anxious or in pain.

      Control:

      • Continuous sedation with no mandated daily interruption.
    • Blinding
      • Not possible
      • Outcomes were objective (days on ventilator, LOS) → minimal detection bias.
    • Conclusion

      Moderate‑to‑strong – robust randomisation and objective outcomes offset by open‑label design and post‑randomisation exclusions.

    External Validity

    • Population Representativeness
      • Single academic medical ICU;
      • No surgical/trauma or paediatric cases;
      • Sedation target Ramsay Sedation Scale score of 3‑4 (deeper than contemporary target of 0/‑1).
    • Applicability
      • Intervention practical where nursing ratios allow close monitoring.
      • May be challenging in low‑staffed or resource‑limited settings.
    • Conclusion
      • Moderate – readily applicable to adult medical ICUs using continuous infusions, but caution in other populations and in units already practising light sedation.

    Strengths & Limitations

    Strengths

    • Simple Intervention

      • Simple, low‑cost intervention with clear mechanistic rationale.

    • Reasonable Methodology
      • Randomisation, with concealed allocation
    • Separation in Sedative Dosing
      • Quantified sedative and opioid exposure, demonstrating biological separation.

    Limitations

    • Design Limitations
      • Single-centre trial
      • Open label, leading to a potential Hawthorn effect
      • No published power calculation
    • Execution
      • 15 % post‑randomisation exclusions
      • 30 % crossover in controls

    Interpretation / Why This Matters

    • Transformational Trial
      • The Kress trial provided a foundational evidence base supporting the strategy of daily sedation interruption for ICU patients.
      • By reducing excessive sedation, clinicians could shorten mechanical ventilation time and ICU stay without increasing major complications.
    • Paradigm Shift
      • This trial challenged longstanding sedation practices and paved the way for further research on protocolized sedation, sedation holidays, and integrated “awake and breathing” trials.
    • Landmark Trial
      • The concept underpins modern ICU liberation bundles and informs guideline recommendations that prioritise either daily sedation interruption or targeting light sedation from the outset.

    Controversies & Subsequent Evidence

    Summary

    • Clinically Impactful
      • Scheduled daily sedation interruption reduced ventilation by ≈ 2.4 days and ICU stay by ≈ 3.5 days.
    • Sedation Reduction
      • Benzodiazepine and opioid exposure (midazolam strata) fell by roughly half.
    • Safe

      No increase in self‑extubation, haemodynamic instability or mortality.

    • Robust Evidence

      Internal validity moderate‑strong; external validity moderate for adult medical ICUs.

    • Paradigm Change
      • Sparked the shift from deep, continuous sedation to light or interrupted sedation worldwide.

    Conclusion

    • Landmark Trial

      The Kress sedation break trial changed how sedation was delivered worldwide and lead to a series of trials seeking to optimise this new paradigm of lighter sedation

    Further Reading

    Other Trials

    Systematic Review & Meta Analysis

    Observational Studies

    Guidelines

    Overall Takeaway

    By proving that a simple, nurse‑led “wake‑up call” each day hastens liberation from the ventilator without compromising safety, the 2000 Kress trial dismantled the culture of uninterrupted deep sedation and laid the foundation for today’s light‑sedation or daily‑awakening standards embedded in international ICU guidelines.