Skip to main content

Foundational Trials

NONSEDA

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Nonsedation or Light Sedation in Critically Ill, Mechanically Ventilated Patients
  • Acronym: NONSEDA
  • Year: 2020
  • Journal published in: New England Journal of Medicine
  • Citation: Olsen HT, Nedergaard HK, Strøm T, Oxlund J, Wian KA, Ytrebø LM, et al. Nonsedation or light sedation in critically ill, mechanically ventilated patients. N Engl J Med. 2020;382(12):1103-1111.

Context & Rationale

  • Background
    Sedation is frequently used to facilitate invasive mechanical ventilation, but deep or prolonged sedation has been linked to delayed liberation, delirium, and longer ICU stays; as a result, sedation minimisation strategies have been tested (e.g., daily interruption of sedative infusions1 and bundled spontaneous awakening + breathing trials2).
    A prior single-centre randomised trial of a “no sedation” protocol suggested potential improvements in resource outcomes, but raised questions about feasibility, safety, and generalisability3.
    Contemporary international guidance increasingly favoured analgesia-first and light sedation targets rather than routine deep sedation4.
  • Research Question/Hypothesis
    In critically ill adults expected to require prolonged ventilation, does a protocol of non-sedation (with analgesia-first care and rescue sedation as needed) reduce 90-day all-cause mortality compared with protocolised light sedation including a daily wake-up trial?
  • Why This Matters
    “No sedation” represents a maximal de-escalation strategy beyond light sedation; if effective and safe, it could reduce exposure to sedatives and downstream complications, but if unsafe it risks device-related harms (e.g., unplanned extubation) and staff burden.
    Mortality and patient-centred outcomes had remained uncertain in modern ICUs where light sedation is already common.

Design & Methods

  • Research Question: Whether a non-sedation strategy improves 90-day survival versus protocolised light sedation with daily wake-up in mechanically ventilated ICU patients.
  • Study Type: Multicentre, investigator-initiated, pragmatic, parallel-group randomised controlled trial in 8 Scandinavian ICUs; open-label clinical care with objective outcomes; stratified randomisation (centre, age category, shock at admission).
  • Population:
    • Setting: Adult ICUs in Denmark, Norway, and Sweden; enrolment early after initiation of invasive ventilation.
    • Key inclusion: Adults receiving invasive mechanical ventilation; screened within 24 hours of intubation; expected mechanical ventilation >24 hours.
    • Key exclusions: Conditions mandating continuous sedation or neurological control (e.g., therapeutic hypothermia, status epilepticus, elevated intracranial pressure/need for controlled ventilation), and severe hypoxaemia (PaO2/FiO2 ≤ 9 kPa in the published protocol).
  • Intervention:
    • Non-sedation strategy: No continuous sedative infusions; analgesia-first (typically opioid boluses/infusion per protocol) with environmental/comfort measures; sedation permitted as rescue (e.g., severe agitation threatening safety or ventilation synchrony).
    • Target state: Awake or lightly sedated when feasible, with mobilisation and delirium prevention measures where possible.
  • Comparison:
    • Light sedation strategy: Protocolised sedation with daily wake-up trial; sedative choice per protocol (propofol during initial period, with midazolam as needed thereafter), titrated to a light sedation target (RASS approximately −2 to −3) and interrupted daily when clinically appropriate.
    • Co-interventions: Both groups received usual ICU care including analgesia, ventilator management, and delirium monitoring/treatment per local protocols.
  • Blinding: No blinding of treating clinicians or patients (behavioural intervention); mortality was objective and captured from registries/records; data adjudication/analysis followed a pre-specified plan.
  • Statistics: Protocol planned n≈700 to detect a 25% relative risk reduction in mortality with 80% power (β=0.20) at a two-sided 5% significance level (α=0.05); primary analysis was modified intention-to-treat5.
  • Follow-Up Period: Primary outcome at 90 days; key secondary outcomes assessed over 28 days (e.g., ventilator-free days, ICU-free days, delirium/coma metrics) and selected safety events during ICU stay.

Key Results

This trial was not stopped early. It completed recruitment and analysed 700 patients in a modified intention-to-treat population (349 non-sedation; 351 light sedation).

Outcome Non-sedation Light sedation Effect p value / 95% CI Notes
All-cause mortality (90 days) (primary) 148/349 (42.4%) 130/351 (37.0%) Risk difference +5.4 percentage points 95% CI −2.2 to 12.2; P=0.65 Modified ITT; CI includes clinically important benefit and harm.
All-cause mortality (28 days) (exploratory) 113/349 (32.4%) 99/351 (28.2%) Risk difference +4.2 percentage points 95% CI −2.6 to 11.0 Exploratory secondary analysis.
Days without mechanical ventilation within 28 days (median, IQR) 20 (0–26) 19 (0–25) Median difference +1 day 95% CI −3 to 3 Ventilator-free days; no multiplicity adjustment.
Days in ICU within 28 days (median, IQR) 13 (0–23) 14 (0–23) Median difference −1 day 95% CI −7 to 4 ICU-free days not improved.
Days free from coma or delirium within 28 days (median, IQR) 27 (21–28) 26 (22–28) Median difference +1 day 95% CI 0 to 2 Composite of coma and delirium assessments.
Major thromboembolic event (0–28 days) 1/349 (0.3%) 10/351 (2.8%) Risk difference −2.5 percentage points 95% CI −4.8 to −0.7 Secondary outcome; interpret cautiously given multiple comparisons.
Accidental self-extubation requiring re-intubation within 24 hours 31/349 (8.9%) 14/351 (4.0%) Risk difference +4.9 percentage points 95% CI 1.3 to 8.7; P=0.01 Safety outcome from trial supplementary appendix.
Accidental removal of “other equipment” (e.g., lines/tubes) 53/349 (15.2%) 32/351 (9.1%) Risk difference +6.1 percentage points 95% CI 1.3 to 11.0; P=0.01 Safety outcome from trial supplementary appendix.
  • Non-sedation did not improve 90-day mortality (42.4% vs 37.0%; risk difference +5.4 percentage points; 95% CI −2.2 to 12.2; P=0.65).
  • Protocol separation existed but was incomplete: mean RASS ranged from −1.3 to −0.8 (non-sedation) vs −2.3 to −1.8 (light sedation) over days 1–7, and 27% of the non-sedation group received sedation within the first 24 hours.
  • Device-related harms were more frequent with non-sedation (e.g., re-intubation after self-extubation within 24 hours: 8.9% vs 4.0%; P=0.01).

Internal Validity

  • Randomisation and allocation: Centralised, concealed allocation with stratification by centre, age category, and shock at admission (pre-specified; implemented early after intubation).
  • Dropout/exclusions: 710 randomised; 10 excluded from the modified ITT (5 per group); primary outcome ascertainment was otherwise complete via records/registries.
  • Performance/detection bias: Open-label care; primary endpoint (all-cause mortality) is objective and less susceptible to ascertainment bias; several secondary outcomes (e.g., delirium assessments, adverse events) are more vulnerable to measurement and care-process differences.
  • Protocol adherence: Clinical separation was measurable but imperfect; rescue sedation occurred in a substantial minority of the non-sedation arm (sedation within 24 hours: 27% non-sedation; any time during ICU stay: 38.4% non-sedation).
  • Baseline characteristics: Groups were broadly comparable; small differences included APACHE II score (median 26 non-sedation vs 25 light sedation) and age (median 72 vs 70 years).
  • Timing: Randomisation within 24 hours of intubation supports evaluation of “early” sedation strategy; later sedation exposure still occurred (notably in the non-sedation arm).
  • Dose/separation of exposure: Mean RASS separation over days 1–7 (approximately 1 RASS unit); sedative exposure differed (e.g., median propofol dose during days 1–2: 0.22 vs 0.84 mg/kg/h; midazolam days 3–7: 0.00 vs 0.03 mg/kg/h), while opioid exposure was similar (morphine days 1–7: 0.20 vs 0.15 mg/kg/day).
  • Outcome assessment: Mortality captured from registries/records; secondary outcomes included ICU process measures and adverse events captured during ICU admission.
  • Statistical rigour: Primary analysis was modified ITT; the trial reported multiple secondary outcomes without adjustment for multiplicity (interpret secondary findings accordingly).

Conclusion on Internal Validity: Overall, internal validity is moderate: randomisation and follow-up for mortality were robust, but the necessarily unblinded design, incomplete separation (rescue sedation), and multiple secondary endpoints without multiplicity control complicate causal inference beyond the primary mortality outcome.

External Validity

  • Population representativeness: Mechanically ventilated adults in well-resourced Scandinavian ICUs (older median age ~70–72); patients with several conditions mandating deep sedation or neurological control were excluded.
  • Setting and resources: Feasibility of non-sedation may depend on staffing and culture; participating ICUs reported high nurse:patient ratios and established protocols.
  • Applicability: Findings are most applicable to systems already using protocolised light sedation; generalisability to ICUs with lower staffing, different delirium management practices, or different sedation norms may be limited.

Conclusion on External Validity: Generalisability is moderate for similar high-resource ICUs practising protocolised light sedation; translation to lower-resource or differently staffed settings is more uncertain.

Strengths & Limitations

  • Strengths: Large, pragmatic multicentre design; early randomisation; objective primary endpoint at 90 days; clear protocol distinction; detailed safety reporting.
  • Limitations: Unblinded behavioural intervention; incomplete separation due to rescue sedation; primary effect estimate compatible with benefit or harm; multiple secondary outcomes increase false-positive risk; resource-intensive setting may limit transferability.

Interpretation & Why It Matters

  • Clinical signal
    Compared with protocolised light sedation, routine non-sedation did not improve mortality and was associated with more device-related adverse events (notably self-extubation requiring re-intubation).
  • Practical implication
    For most ICUs, the trial supports maintaining an analgesia-first, light sedation strategy rather than adopting a default “no sedation” approach for all ventilated patients.
  • Mechanistic takeaway
    In settings already targeting light sedation, incremental reduction towards “no sedation” may yield diminishing returns and increase safety risks without clear patient-centred benefit.

Controversies & Subsequent Evidence

  • Power and detectable effect size: The trial was designed to detect a large mortality effect; the observed CI still allows for smaller benefit or harm, meaning “no difference” should be read as “no evidence of a large benefit” rather than proof of equivalence6.
  • Exposure separation and pragmatic rescue sedation: Rescue sedation was common in the non-sedation group, reflecting real-world feasibility constraints and potentially diluting treatment contrast; an in-depth post hoc assessment characterised “failures” of the non-sedation strategy and their drivers in routine care10.
  • Safety trade-offs: Increased unplanned device-related events with non-sedation (self-extubation, removal of other equipment) underscores the importance of staffing, mobilisation policies, delirium prevention, and rescue pathways when sedation is minimised.
  • Patient-centred survivorship outcomes: A subsequent follow-up analysis of survivors did not demonstrate clear longer-term functional advantage attributable to a default non-sedation strategy, despite potential early differences in wakefulness/mobilisation9.
  • How the wider evidence base moved: Updated guideline work continues to emphasise analgesia-first care and light sedation targets while warning against over-sedation; recent SCCM-focused updates maintain this trajectory7.
  • Synthesis of randomised evidence: Modern systematic reviews/meta-analyses evaluating sedation depth and protocolised sedation strategies suggest limited or no mortality effect from incremental sedation reductions in contemporary practice, and increasingly focus on safety and process outcomes (e.g., ventilation duration, delirium, self-extubation) rather than expecting large survival gains8.

Summary

  • NONSEDA compared a protocolised non-sedation strategy versus protocolised light sedation with daily wake-up in 700 mechanically ventilated ICU patients.
  • There was no mortality benefit at 90 days (42.4% non-sedation vs 37.0% light sedation; risk difference +5.4 percentage points; 95% CI −2.2 to 12.2; P=0.65).
  • Exposure separation was modest and incomplete (mean RASS day 1–7 ≈ −1.3 to −0.8 vs −2.3 to −1.8; rescue sedation within 24 hours in 27% of the non-sedation group).
  • Non-sedation was associated with more device-related harms (self-extubation requiring re-intubation within 24 hours: 8.9% vs 4.0%; P=0.01).
  • The trial supports a pragmatic default of analgesia-first, light sedation rather than routine “no sedation” for all ventilated patients in modern ICUs.

Further Reading

Other Trials

Systematic Review & Meta Analysis

Observational Studies

Guidelines

Notes

  • NONSEDA is best interpreted as “non-sedation vs protocolised light sedation” in a system with established sedation protocols and high staffing; it does not test non-sedation against routine deep sedation practice.

Overall Takeaway

NONSEDA is a landmark pragmatic test of whether pushing beyond light sedation to a default non-sedation strategy improves survival in modern ventilated ICU care. It did not reduce 90-day mortality and it increased important device-related harms, supporting contemporary practice that prioritises analgesia-first and light (rather than absent) sedation for most mechanically ventilated adults.

Overall Summary

  • No mortality benefit with routine non-sedation versus protocolised light sedation.
  • Meaningful but incomplete separation; rescue sedation was frequent in the non-sedation arm.
  • Non-sedation increased device-related adverse events (notably self-extubation requiring re-intubation).

Bibliography