LandmarkRCT

ALVEOLI

Higher versus Lower Positive End-Expiratory Pressures in Patients with the Acute Respiratory Distress Syndrome

New England Journal of Medicine, 2004 · Multicenter, randomized controlled trial (ARDSNet), unblinded, stopped early for futility.

Clinical question

Among patients already ventilated with a low-tidal-volume, plateau-pressure-limited strategy, does a higher-PEEP/lower-FiO2 approach improve survival compared with a lower-PEEP/higher-FiO2 approach?

Population: 549 patients with acute lung injury or ARDS, all ventilated with 6 mL/kg predicted body weight and a plateau pressure ceiling of 30 cmH2O in both arms.
Inclusion criteria: Acute lung injury or ARDS by the contemporaneous consensus definition.
Enrolled from ARDS Network centers, on low tidal volume ventilation in both study arms.
Intervention: Higher-PEEP, lower-FiO2 strategy, using a PEEP/FiO2 combination table weighted toward higher PEEP.
Comparator: Lower-PEEP, higher-FiO2 strategy, using a PEEP/FiO2 table weighted toward lower PEEP.
Primary outcome: In-hospital mortality before discharge home and unassisted breathing.
Secondary outcomes: Ventilator-free days through day 28.
Result: No difference in mortality or ventilator-free days between higher- and lower-PEEP strategies when both arms used low tidal volume ventilation. (Ventilator-free days 13.8 ± 10.6 (higher PEEP) vs 14.5 ± 10.4 (lower PEEP), p = 0.50)

What the trial found

The trial was stopped early for futility: a higher-PEEP strategy produced clinical outcomes statistically indistinguishable from a lower-PEEP strategy once both groups were already receiving lung-protective tidal volumes. It did not show that higher PEEP is harmful, nor that it is beneficial on average.

Current guideline position

The 2023 ESICM ARDS guideline conditionally favors higher PEEP without a recruitment maneuver over lower PEEP specifically in moderate-to-severe ARDS (low certainty) — a narrower and more recent statement than this trial's own null result in an all-comers population.

Clinical interpretation

ALVEOLI does not by itself justify routinely using either a higher- or a lower-PEEP protocol — it shows the two fixed PEEP/FiO2 tables tested performed similarly on average across an unselected ALI/ARDS population. Later individual-patient-data meta-analyses (not reproduced here) suggested a mortality signal favoring higher PEEP specifically in moderate-to-severe disease, which is why current guidance is severity-stratified rather than a flat 'higher is better' or 'no difference' statement.

Limitations

Stopped early for futility, which reduces power to detect a smaller true effect.
Used fixed PEEP/FiO2 tables rather than individualized titration (e.g. by esophageal pressure or driving pressure), which may not reflect how PEEP is actually selected at the bedside.
Enrolled a mixed acute-lung-injury/ARDS population without severity stratification, which can dilute a subgroup-specific effect.

Moderate certainty

Clinical Context & Relevant Topics (1)

Disease Reference

Acute Respiratory Distress Syndrome

Acute, diffuse inflammatory lung injury causing hypoxemia and bilateral infiltrates not fully explained by cardiac failure or fluid overload.

View Topic Context →