Skip to main content

Foundational Trials

NIV for AECOPD

Image: Shutterstock / ChaNaWiT

Publication

  • Title: Noninvasive Ventilation for Acute Exacerbations of Chronic Obstructive Pulmonary Disease
  • Acronym: Not applicable
  • Year: 1995
  • Journal published in: New England Journal of Medicine
  • Citation: Brochard L, Mancebo J, Wysocki M, Lofaso F, Conti G, Rauss A, Simonneau G, Benito S, Gasparetto A, Lemaire F, Isabey D, Harf A. Noninvasive ventilation for acute exacerbations of chronic obstructive pulmonary disease. N Engl J Med. 1995 Sep 28;333(13):817-822.

Context & Rationale

  • Background
    • Acute exacerbations of COPD (AECOPD) complicated by acute hypercapnic respiratory acidosis frequently prompt ICU admission and endotracheal intubation.
    • Invasive mechanical ventilation can be life-saving but exposes patients to sedation-related complications, ventilator-associated infections, barotrauma, and prolonged weaning trajectories in a population prone to dynamic hyperinflation and difficult liberation.
    • Non-invasive ventilation (NIV) via a mask was physiologically plausible as a means to offload respiratory muscles and correct hypercapnia while avoiding airway instrumentation, but robust evidence for hard patient-centred outcomes in ICU-treated AECOPD was limited.
  • Research Question/Hypothesis
    • In ICU patients with AECOPD and respiratory acidosis who did not require immediate intubation, would early face-mask pressure-support NIV in addition to standard medical therapy reduce the need for endotracheal intubation and improve clinical outcomes compared with standard therapy alone?
  • Why This Matters
    • Demonstrating that NIV prevents intubation would plausibly reduce ventilation-related complications, shorten hospitalisation, and lower mortality in a high-risk, high-resource-use ICU cohort.
    • Positive results could reframe ventilatory escalation pathways for AECOPD and shift practice towards earlier, mask-based ventilatory support with explicit failure criteria.

Design & Methods

  • Research Question: In ICU patients with AECOPD and respiratory acidosis not requiring immediate intubation, does early face-mask pressure-support NIV plus standard therapy reduce the need for endotracheal intubation compared with standard therapy alone?
  • Study Type: Multicentre, randomised, parallel-group clinical trial; ICU setting across 5 centres in 3 European countries (France, Italy, Spain); unblinded.
  • Population:
    • Setting: ICU admissions for AECOPD (screened 275 admissions; enrolled 85).
    • Key inclusion features: Known COPD (or high probability based on history/clinical exam/chest radiograph); exacerbation of dyspnoea <2 weeks; respiratory acidosis with elevated bicarbonate; and at least 2 of: respiratory rate >30 breaths/min, PaO2 <45 mmHg, arterial pH <7.35 (after 10 min breathing room air).
    • Key exclusions: Immediate intubation requirement (respiratory arrest; severe haemodynamic instability; inability to tolerate mask; impaired consciousness; airway clearance issues; high aspiration risk; recent upper GI surgery); respiratory rate <12; intubation before admission; tracheostomy; sedative drugs within 12 h; non-hypercapnic CNS disorder; cardiac arrest within 5 days; cardiogenic pulmonary oedema; kyphoscoliosis/neuromuscular disease; upper airway obstruction/asthma; clear alternative cause requiring specific treatment (e.g., septic shock, acute MI, pneumothorax, pulmonary embolism, severe pneumonia, recent surgery/trauma); facial deformity; refusal of intubation.
  • Intervention:
    • NIV mode/interface: Pressure-support ventilation delivered via face mask.
    • Initial settings: Pressure support 20 cm H2O, adjusted to achieve maximal patient comfort; expiratory pressure set at atmospheric pressure (no applied PEEP).
    • Oxygen: Supplemented during NIV to maintain arterial oxygen saturation >90%.
    • Delivery schedule: At least 6 h/day; up to 2 h/day allowed for spontaneous breathing; duration determined by clinical and arterial blood gas criteria.
    • Co-interventions: Standard medical therapy in both arms (antibiotics when indicated, bronchodilators, corticosteroids, anticoagulation, electrolyte correction, and additional agents as per protocolised ICU care).
  • Comparison:
    • Standard therapy: Oxygen via nasal prongs (max 5 L/min) titrated to maintain arterial oxygen saturation >90% plus protocolised medical therapy (as above).
    • Intubation approach: Predefined intubation criteria applied; intubation performed if criteria were met after 1 h of assigned study treatment (additional criteria applied in the NIV arm following withdrawal of NIV, including a step of NIV resumption prior to intubation if criteria were met).
  • Blinding: Unblinded; no sham ventilation; intubation criteria were protocolised to mitigate clinician discretion.
  • Statistics: Power calculation (effect size, alpha, beta/power, required sample size): Not reported; hypothesis testing used two-tailed P values with P<0.05 considered significant; analyses presented by randomised group (intention-to-treat not explicitly stated).
  • Follow-Up Period: In-hospital outcomes through discharge or death; spirometry and arterial blood gases repeated within 3 months after hospital discharge.

Key Results

This trial was not stopped early.

Outcome NIV + standard therapy (N=43) Standard therapy (N=42) Effect p value / 95% CI Notes
Need for endotracheal intubation (primary outcome) 11/43 (26%) 31/42 (74%) RR/OR not reported P<0.001; 95% CI not reported Most intubations occurred within 12 h (9/11 vs 23/31); intubation criteria were predefined.
In-hospital mortality 4/43 (9%) 12/42 (29%) RR/OR not reported P=0.02; 95% CI not reported Difference in mortality was reported as not significant after adjustment for endotracheal intubation (Mantel–Haenszel).
Patients with ≥1 complication not present on admission 7/43 (16%) 20/42 (48%) RR/OR not reported P=0.001; 95% CI not reported Total complications: 9 vs 23; complications leading to death: 4 vs 12.
Length of hospital stay (days) 23 ± 17 35 ± 33 Mean difference not reported P=0.02; 95% CI not reported Subgroup (post-randomisation): no-intubation 17 ± 9 vs 20 ± 16; intubated 40 ± 22 vs 41 ± 36.
Interface complication: facial-skin necrosis (event count) 1 0 Not reported Not reported Reported among complications not present on admission; not associated with death.
  • Marked between-group separation in ventilatory escalation: intubation 11/43 (26%) with NIV vs 31/42 (74%) with standard therapy.
  • Rapid physiological response consistent with reduced work of breathing and improved ventilation at 1 h: respiratory rate 25 ± 8 vs 33 ± 7 breaths/min; PaCO2 55 ± 9 vs 70 ± 15 mmHg; pH 7.33 ± 0.04 vs 7.27 ± 0.06 (NIV vs standard).
  • Clinical outcomes aligned with intubation avoidance: fewer patients with complications (7/43 vs 20/42) and shorter hospitalisation (23 ± 17 vs 35 ± 33 days).

Internal Validity

  • Randomisation and Allocation:
    • Randomisation performed using sealed envelopes; additional safeguards for allocation concealment were not reported.
    • Allocation ratio: 43 assigned to NIV + standard therapy vs 42 to standard therapy.
  • Drop out or exclusions:
    • Post-randomisation exclusions from the primary clinical outcomes: Not reported.
    • Early intubation reduced the 1 h physiological dataset: standard therapy N=39 (of 42) and NIV N=42 (of 43) for 1 h measurements.
  • Performance/Detection Bias:
    • Unblinded intervention; however, endotracheal intubation criteria were predefined and applied during the study.
    • Secondary outcomes (complications and length of stay) include elements influenced by clinician behaviour and institutional practice patterns.
  • Protocol Adherence:
    • NIV delivery protocol required ≥6 h/day with pressure support initiated at 20 cm H2O; actual delivered hours/day and settings distribution were not reported.
    • Standard therapy specified oxygen via nasal prongs up to 5 L/min targeting arterial saturation >90%.
  • Baseline Characteristics:
    • Groups were similar at randomisation across key severity markers: pH 7.28 ± 0.06 (standard) vs 7.27 ± 0.06 (NIV); PaCO2 68 ± 15 vs 69 ± 13 mmHg; respiratory rate 35 ± 6 vs 37 ± 8 breaths/min; SAPS 11 ± 5 vs 11 ± 4.
  • Heterogeneity:
    • Multicentre conduct (5 ICUs) introduces centre-level variation; intubation frequencies varied by centre but the overall between-group difference persisted (total 31/42 vs 11/43).
  • Timing:
    • Early effect assessment at 1 h; most endotracheal intubations occurred early (within 12 h: 23/31 in standard vs 9/11 in NIV).
  • Dose:
    • Pressure-support NIV with initial inspiratory assistance of 20 cm H2O; no applied expiratory pressure (atmospheric expiratory pressure).
  • Separation of the Variable of Interest:
    • Endotracheal intubation: 11/43 (26%) vs 31/42 (74%).
    • At 1 h: PaCO2 55 ± 9 vs 70 ± 15 mmHg; pH 7.33 ± 0.04 vs 7.27 ± 0.06; respiratory rate 25 ± 8 vs 33 ± 7 (NIV vs standard).
  • Outcome Assessment:
    • Primary endpoint (intubation) is clinically concrete; protocol specified explicit intubation triggers.
    • Complications were defined as not present on admission; event adjudication process was not reported.
  • Statistical Rigor:
    • Two-tailed hypothesis testing with P<0.05 as significance threshold; multiple tests reported across physiological and clinical outcomes.
    • Pre-specified sample size/power calculation was not reported.

Conclusion on Internal Validity: Overall, internal validity appears moderate-to-strong given randomisation, baseline balance, protocolised intubation criteria, and large between-group separation in clinically meaningful outcomes, tempered by unblinded care and limited reporting of allocation safeguards and intervention fidelity.

External Validity

  • Population Representativeness:
    • Highly selected ICU cohort: 85 enrolled from 275 screened admissions; exclusions included immediate intubation needs and numerous alternate diagnoses (e.g., severe pneumonia, septic shock, acute MI, pneumothorax).
    • Physiological severity at enrolment (room air PaO2 ~42–43 mmHg; pH ~7.27–7.28; PaCO2 ~68–69 mmHg) reflects a specific subset of AECOPD with established respiratory acidosis.
  • Applicability:
    • Conducted in monitored ICU environments with access to urgent intubation; applicability to unmonitored wards depends on staffing, training, and escalation pathways.
    • NIV settings (pressure support without applied expiratory pressure) and oxygen targets (>90%) differ from many contemporary bilevel NIV protocols and controlled oxygen strategies, but the core physiological principle (augment alveolar ventilation) remains relevant.

Conclusion on External Validity: Generalisability is moderate for patients with AECOPD and acute hypercapnic respiratory acidosis who do not require immediate intubation and can be managed in a closely monitored setting; it is more limited for patients with mixed aetiologies (e.g., pneumonia/sepsis), those needing immediate airway control, or settings without reliable NIV expertise and escalation capacity.

Strengths & Limitations

  • Strengths:
    • Multicentre randomised design across several ICUs and countries.
    • Clinically decisive primary endpoint (need for endotracheal intubation) with protocolised criteria.
    • Substantial and rapid physiological separation consistent with the mechanism of action (ventilatory unloading and CO2 clearance).
    • Patient-centred secondary outcomes (complications, mortality, length of stay) captured during the index hospitalisation.
  • Limitations:
    • Unblinded delivery with potential influence on discretionary outcomes (e.g., length of stay, complication recognition).
    • Small sample size with no reported pre-specified power calculation.
    • Highly selected ICU population (minority of screened admissions), potentially limiting generalisability and inflating apparent effect size relative to less severe cohorts.
    • Ventilation strategy differs from many contemporary bilevel NIV protocols (no applied expiratory pressure/PEEP).

Interpretation & Why It Matters

  • Clinical implication
    In selected AECOPD with respiratory acidosis (pH <7.35) and without immediate intubation requirements, early NIV can avert progression to invasive ventilation and its attendant complications.
  • Mechanistic coherence
    Rapid improvement in PaCO2, pH, and respiratory rate within 1 h supports a causal pathway from ventilatory unloading to reduced need for intubation and downstream reductions in complications and hospital stay.
  • Systems and safety
    NIV benefit depends on early initiation, skilled bedside delivery, frequent reassessment, and explicit failure criteria to ensure timely transition to invasive ventilation when required.

Controversies & Other Evidence

Summary

  • In 85 ICU patients with AECOPD and respiratory acidosis, early face-mask pressure-support NIV plus standard therapy reduced endotracheal intubation (11/43 vs 31/42; P<0.001).
  • NIV was associated with fewer patients developing new complications (7/43 vs 20/42; P=0.001) and shorter hospital stay (23 ± 17 vs 35 ± 33 days; P=0.02).
  • In-hospital mortality was lower in the NIV group (4/43 vs 12/42; P=0.02), with reported attenuation after adjustment for intubation, supporting a mediation pathway through intubation avoidance.
  • Physiological improvements were evident within 1 h (PaCO2 55 ± 9 vs 70 ± 15 mmHg; pH 7.33 ± 0.04 vs 7.27 ± 0.06; NIV vs standard).
  • This trial helped establish NIV as first-line ventilatory support for acute hypercapnic AECOPD in appropriately selected patients, contingent on skilled delivery and timely escalation when NIV fails.

Overall Takeaway

Brochard and colleagues demonstrated that early non-invasive pressure-support ventilation via face mask in selected ICU patients with AECOPD and respiratory acidosis substantially reduced the need for endotracheal intubation and was associated with fewer complications, shorter hospitalisation, and lower in-hospital mortality. The trial’s enduring impact is the modern practice norm that NIV is first-line ventilatory support for acute hypercapnic COPD exacerbations, provided systems reliably deliver early NIV, close monitoring, and clear escalation to invasive ventilation when necessary.

Overall Summary

  • Early NIV in selected AECOPD with respiratory acidosis reduced invasive ventilation requirements and downstream morbidity, underpinning contemporary guideline-endorsed first-line use with robust failure/escalation pathways.

Bibliography