Skip to main content

Foundational Trials

IRIS

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Effect of Cricoid Pressure Compared With a Sham Procedure in the Rapid Sequence Induction of Anesthesia: The IRIS Randomized Clinical Trial
  • Acronym: IRIS (Sellick Interest in Rapid Sequence Induction)
  • Year: 2019
  • Journal published in: JAMA Surgery
  • Citation: Birenbaum A, Hajage D, Roche S, et al. Effect of cricoid pressure compared with a sham procedure in the rapid sequence induction of anesthesia: The IRIS randomized clinical trial. JAMA Surg. 2019;154(1):9-17.

Context & Rationale

  • Background
    • Cricoid pressure (Sellick manoeuvre) has been widely incorporated into rapid sequence induction (RSI) as an aspiration-prevention manoeuvre, despite limited direct trial evidence linking it to reduced pulmonary aspiration.
    • Physiologic and imaging studies have raised concern that cricoid pressure can distort upper airway anatomy (including postcricoid hypopharyngeal compression), potentially worsening laryngoscopic view and prolonging intubation.
    • The aspiration endpoint is rare in modern practice; therefore, prior studies were underpowered to address clinically important outcomes (aspiration/aspiration pneumonia) rather than surrogate endpoints.
  • Research Question/Hypothesis
    • Whether omitting cricoid pressure (sham procedure) is non-inferior to cricoid pressure during RSI for preventing pulmonary aspiration in surgical patients with a “full stomach” or other aspiration risk factors.
  • Why This Matters
    • Cricoid pressure sits at the intersection of low-frequency catastrophic harm (aspiration) and high-frequency process outcomes (laryngoscopic view, intubation difficulty); a large pragmatic RCT was needed to inform practice.
    • Noninferiority framing was clinically salient: if sham could be shown non-inferior, routine cricoid pressure could be reconsidered, potentially improving airway conditions without increasing aspiration.

Design & Methods

  • Research Question: In adult surgical patients undergoing RSI with aspiration risk factors, is a sham procedure non-inferior to cricoid pressure (Sellick manoeuvre) for preventing pulmonary aspiration?
  • Study Type: Randomised, multicentre (10 academic centres), double-blind, non-inferiority, parallel-group trial (France); enrolment February 2014 to February 2017; follow-up to 28 days or hospital discharge.
  • Population:
    • Adults undergoing surgery under general anaesthesia requiring RSI.
    • “Full stomach” defined as <6 hours fasting.
    • Or ≥1 aspiration risk factor: gastro-oesophageal reflux; diabetes; opioid use; peptic ulcer; previous gastric surgery; obesity (BMI >30); pregnancy; intestinal obstruction; delayed gastric emptying.
    • Key exclusions included: inability to provide informed consent (notably, emergency RSI in patients unable to consent); outpatient surgery; RSI without tracheal intubation.
  • Intervention:
    • Cricoid pressure (Sellick manoeuvre) applied with target force 30 N using the first three fingers on the cricoid cartilage.
    • Operator training used a 50 mL syringe model (compressing from 40 mL to 33 mL to approximate 30 N) to standardise applied force.
    • Applied by an anaesthesiologist not otherwise involved in patient care; a screen/cover was used to blind the intubating clinician to manoeuvre allocation.
  • Comparison:
    • Sham manoeuvre with identical positioning/ritualisation but without applying cricoid pressure.
    • Same blinding approach (screen/cover) to preserve clinician blinding.
    • Cricoid pressure/sham could be interrupted if requested by the intubating clinician (e.g., impaired view).
  • Blinding: Double-blind for the anaesthesiologist and intubating clinician (screen/cover over neck); the individual applying the manoeuvre was not blinded but was not otherwise involved in airway management.
  • Statistics: A total of 3434 patients (1717 per arm) were required to test non-inferiority of sham vs Sellick for pulmonary aspiration, assuming 1% event incidence and a non-inferiority margin corresponding to a 50% relative increase (RR margin 1.5), with 80% power and a one-sided alpha of 2.5% (reported as 0.025). Target sample size was 3500 allowing 1.5% consent withdrawal; primary analysis was per-protocol with supportive intention-to-treat analysis.
  • Follow-Up Period: Primary endpoint assessed peri-intubation; secondary outcomes included early pulmonary outcomes (first 24 hours and first 7 days) and mortality to day 28 (or hospital discharge).

Key Results

This trial was not stopped early. Recruitment continued to the planned sample size (3472 randomised across 10 centres).

Outcome Sellick (cricoid pressure) Sham Effect p value / 95% CI Notes
Pulmonary aspiration (primary; per-protocol) 10/1729 (0.6%) 9/1730 (0.5%) RR 0.90 90% CI 0.39 to 2.00; P=0.14 RR expressed as Sham/Sellick; upper CI (2.00) exceeded non-inferiority margin (1.50); non-inferiority not demonstrated.
Pulmonary aspiration (primary; intention-to-treat) 10/1784 (0.6%) 9/1736 (0.5%) RR 0.90 90% CI 0.39 to 1.99; P=0.14 Risk difference (Sham − Sellick) −0.06% (two-sided 95% CI −0.57 to 0.42); non-inferiority not demonstrated.
Aspiration pneumonia (first 24 hours) 16/1735 (0.9%) 15/1736 (0.9%) OR 0.94 95% CI 0.46 to 1.93; P=0.86 Secondary endpoint; odds ratios reported in trial publication.
Severe aspiration pneumonia (first 7 days) 5/1735 (0.3%) 7/1736 (0.4%) OR 1.40 95% CI 0.45 to 4.40; P=0.56 Secondary endpoint; wide CI due to rarity.
Suspected aspiration pneumonia (first 24 hours) 160/1735 (9.2%) 169/1736 (9.7%) OR 1.06 95% CI 0.84 to 1.35; P=0.61 Defined as new infiltrate on chest radiography within 24 hours.
Mortality (day 28) 30/1735 (1.7%) 27/1736 (1.6%) OR 0.90 95% CI 0.53 to 1.54; P=0.70 Follow-up to day 28 or discharge.
Cormack–Lehane grade 3–4 (first attempt) 175/1736 (10%) 93/1736 (5%) Not reported P<0.001 Worse laryngoscopic view with cricoid pressure.
Intubation time, median (IQR) 27 (20–38) s 23 (18–32) s Not reported P<0.001 Prolonged intubation time with cricoid pressure.
Traumatic complications 47/1736 (3%) 25/1736 (1%) Not reported P<0.001 Includes dental damage (0.6% vs 0.2%; P=0.03) and other recorded trauma events.
  • Sham could not be declared non-inferior to Sellick for pulmonary aspiration because the upper confidence limit for RR (2.00) exceeded the pre-specified non-inferiority margin (1.50), despite very low absolute aspiration rates (~0.5–0.6%).
  • Cricoid pressure meaningfully worsened airway process metrics (Cormack–Lehane grade 3–4: 10% vs 5%; P<0.001) and prolonged intubation time (median 27 vs 23 seconds; P<0.001).
  • Cricoid pressure was interrupted more often (14% vs 5%; P<0.001) and was associated with more traumatic complications (3% vs 1%; P<0.001).

Internal Validity

  • Randomisation and allocation concealment: Centralised web-based randomisation (block size 6), stratified by centre; allocation concealment maintained until assignment.
  • Blinding (performance/detection bias): Intubating clinician and managing anaesthesiologist were blinded using a neck screen/cover; the manoeuvre performer was unblinded but not otherwise involved in airway management.
  • Dropout / exclusions: 3472 randomised; intention-to-treat population 1735 vs 1736; per-protocol population 1729 vs 1730; major protocol violations were uncommon (12/3471; 0.3%).
  • Follow-up completeness: Primary endpoint had no missing values; for day-28 outcomes, losses to follow-up were 33/1735 (1.9%) vs 28/1736 (1.6%).
  • Protocol adherence and separation: Substantial “release” of the assigned manoeuvre occurred by clinical request: interruption of cricoid pressure/sham 246/1736 (14%) vs 86/1736 (5%); P<0.001.
  • Separation of the variable of interest (numeric): Evidence of clinically meaningful separation and mechanistic plausibility for airway difficulty: Cormack–Lehane grade 3–4 was 175/1736 (10%) with Sellick vs 93/1736 (5%) with sham; P<0.001; median intubation time was 27 (20–38) s vs 23 (18–32) s; P<0.001.
  • Event rate and power: Pulmonary aspiration incidence was much lower than the design assumption (observed ~0.5–0.6%); this widened confidence intervals and reduced the ability to conclude non-inferiority.
  • Outcome assessment: Primary endpoint (observed pulmonary aspiration) is clinically grounded but remains a low-frequency event with potential ascertainment variability; secondary pneumonia endpoints used radiographic criteria within defined time windows.
  • Statistical rigor: Non-inferiority margin was pre-specified (RR 1.5); primary analysis per-protocol with supportive intention-to-treat analysis; confidence interval framing aligned with the non-inferiority approach (two-sided 90% CI corresponding to one-sided 95% upper bound).

Conclusion on Internal Validity: Moderate-to-strong: randomisation and blinding were methodologically robust and protocol violations were rare, but frequent manoeuvre interruption and a markedly lower-than-anticipated primary event rate limited the inferential strength of the non-inferiority conclusion.

External Validity

  • Population representativeness: Surgical RSI population with defined aspiration risks (including emergency surgery); highest-acuity “crash” intubations and patients unable to consent (a common real-world emergency RSI cohort) were excluded.
  • Setting dependence: Operating theatre environment with anaesthesia-trained teams and standardised application training; transferability to ED/ICU/pre-hospital RSI (different operator mix, time pressure, physiology, and aspiration-risk distributions) is uncertain.
  • Intervention fidelity in practice: Target force (30 N) and formal training may not reflect real-world variability in technique; conversely, frequent interruption suggests that “persistent” cricoid pressure may be uncommon in challenging airways.
  • Outcome transportability: Extremely low aspiration incidence in this trial may not match contexts with higher baseline aspiration risk (e.g., profoundly obtunded emergency presentations), limiting the ability to extrapolate absolute risks.

Conclusion on External Validity: Moderate: findings are most applicable to elective/emergency surgical RSI under anaesthesia care in well-resourced settings; generalisability to ED/ICU RSI and to patients unable to consent is limited.

Strengths & Limitations

  • Strengths: Large multicentre double-blind RCT; clinically meaningful primary endpoint (aspiration); pre-specified non-inferiority design and analysis strategy; low protocol violation rate; objective reporting of airway process metrics and complications.
  • Limitations: Primary event rarer than assumed (reduced precision); non-inferiority not demonstrated (inconclusive rather than equivalence); frequent manoeuvre interruption; exclusion of patients unable to consent (limits applicability to the sickest emergency RSI cohort); inability to directly measure applied force at the bedside beyond training approximation.

Interpretation & Why It Matters

  • Non-inferiority not shown
    Sham could not be declared non-inferior to Sellick for pulmonary aspiration because the confidence interval crossed the non-inferiority margin (upper bound 2.00 vs margin 1.50); this is best interpreted as an inconclusive non-inferiority result rather than proof of benefit for cricoid pressure.
  • Airway mechanics trade-off
    Cricoid pressure worsened laryngoscopic view and prolonged intubation time (Cormack–Lehane grade 3–4: 10% vs 5%; median intubation time: 27 vs 23 seconds), and increased recorded traumatic complications (3% vs 1%), supporting a pragmatic approach where cricoid pressure is readily released when it impairs airway management.
  • Practice implication
    For peri-operative RSI, IRIS supports heightened awareness that cricoid pressure can worsen intubation conditions without establishing a clinically important reduction in aspiration; clinicians need explicit “release criteria” and team training for safe, rapid removal of pressure when airway difficulty emerges.

Controversies & Subsequent Evidence

  • The accompanying commentary argued that the IRIS result should not be interpreted as vindication of cricoid pressure, emphasising long-standing evidentiary fragility, plausible mechanisms of airway compromise, and the need for clinician readiness to release pressure when it impairs laryngoscopy or ventilation.1
  • Non-inferiority was undermined by a lower-than-anticipated aspiration event rate, yielding wide confidence intervals; this creates an interpretive tension between “absence of demonstrated non-inferiority” and the practical reality that absolute aspiration incidence was extremely low in both groups.
  • Mechanistic and imaging work (pre-dating IRIS) has demonstrated that cricoid pressure may not reliably compress the oesophagus as traditionally conceptualised, with compression of postcricoid hypopharyngeal structures occurring instead; these data support biologic plausibility for worsened laryngoscopic conditions observed in IRIS.2
  • Comprehensive reviews have highlighted variability in cricoid pressure technique, uncertainty regarding effectiveness in aspiration prevention, and the potential for harm via airway distortion—contextualising IRIS as a necessary but not definitive trial in an area with persistent practice heterogeneity.3
  • Trauma airway guidance has historically included cricoid pressure among RSI adjuncts, but recommendations have been constrained by evidence quality and balanced against airway management priorities; IRIS reinforces the clinical logic embedded in such guidance that airway management success and oxygenation take precedence when cricoid pressure impairs intubation.4

Summary

  • IRIS was a large, double-blind, multicentre non-inferiority RCT comparing cricoid pressure (Sellick) vs sham during surgical RSI in patients with aspiration risk factors.
  • Pulmonary aspiration was rare (≈0.5–0.6%); sham could not be declared non-inferior to Sellick because the confidence interval exceeded the non-inferiority margin (upper bound 2.00 vs margin 1.50).
  • Cricoid pressure worsened airway conditions and process measures (Cormack–Lehane grade 3–4: 10% vs 5%; median intubation time: 27 vs 23 seconds), and increased manoeuvre interruption (14% vs 5%).
  • Traumatic complications were more frequent with cricoid pressure (3% vs 1%), including higher dental injury (0.6% vs 0.2%).
  • Interpretation is constrained by low primary event rate and frequent release of pressure, but the trial materially informs modern RSI practice by quantifying trade-offs between aspiration prevention uncertainty and airway difficulty.

Further Reading

Other Trials

Systematic Review & Meta Analysis

  • 2017Salem MR, Khorasani A, Saatee S, Crystal GJ, El-Orbany M. Cricoid pressure controversies: Narrative review. Anesthesiology. 2017;126(4):738-752.
  • 2015Algie CM, Mahar RK, Tan HB, Wilson G. Effectiveness and risks of cricoid pressure during rapid sequence induction for endotracheal intubation. Cochrane Database Syst Rev. 2015;(11):CD011656. (DOI not reported in available source text)
  • 2020White L, Cocks TM, Hutton A. Cricoid pressure: Research, practice and education. A scoping review. Heart Lung. 2020;49:87-94. (DOI not reported in available source text)
  • 2021Hung KC, Hung CT, Poon YY, Wu SC, Chen KH, Chen JY. The effect of cricoid pressure on tracheal intubation in adult patients: A systematic review and meta-analysis. Can J Anaesth. 2021;68:137-147. (DOI not reported in available source text)

Observational Studies

Guidelines

Notes

  • Where a DOI is not reported above, it was not available in the provided trial manuscript/editorial reference text supplied for this summary.
  • The IRIS trial publication itself should be consulted for full reference metadata and any updated journal linking.

Overall Takeaway

IRIS is a landmark because it brought rigorous, double-blind, multicentre randomised methodology to a long-entrenched RSI practice that had persisted largely on physiologic rationale and low-level evidence. It did not establish that omitting cricoid pressure is non-inferior for preventing pulmonary aspiration, but it did quantify clinically important airway-management trade-offs (worse laryngoscopic view, longer intubation time, more trauma) that meaningfully inform contemporary RSI decision-making.

Overall Summary

  • Sham could not be declared non-inferior to cricoid pressure for pulmonary aspiration, largely due to imprecision from very low aspiration event rates.
  • Cricoid pressure worsened laryngoscopic conditions and prolonged intubation, with more recorded traumatic complications.
  • Modern RSI practice should incorporate explicit readiness to release cricoid pressure when it impairs ventilation or intubation.

Bibliography