Turbine driven versus compressed gas ventilators for noninvasive ventilation in infants with acute respiratory failure: a randomized crossover trial

Technical Summary: Turbine-Driven vs. Compressed-Gas Ventilators for Noninvasive Ventilation in Infants

Problem Statement
Noninvasive ventilation (NIV) is increasingly utilized for infants with acute respiratory failure across various clinical settings, including pediatric intensive care units (PICUs), emergency departments, and transport medicine. While the efficacy of NIV depends on ventilatory assistance, its success is critically linked to patient–ventilator interaction (synchrony). Asynchrony can impair NIV effectiveness, increase the work of breathing, and prolong hospital stays.

Current NIV delivery relies on diverse technologies and circuit configurations. Conventional PICU ventilators typically employ dual-limb circuits (DLCs) with active expiratory valves and compressed gas sources. In contrast, turbine-driven ventilators often utilize single-limb circuits (SLCs) with intentional leaks (e.g., whisper valves) to manage exhalation and CO₂ washout. Despite the widespread adoption of these different systems, the physiological impact of ventilator technology and circuit configuration on patient–ventilator interaction in infants has not been systematically evaluated. Infants present unique challenges, including small tidal volumes, short neural inspiratory times, and dynamic interface leaks, which complicate trigger detection and cycling performance.

Methodology
The authors conducted a single-center, randomized, crossover physiological study at the Fondazione IRCCS Ca' Granda Ospedale Maggiore Policlinico in Milan, Italy (January 2022 – December 2025).

• Study Population: The study enrolled infants aged 1–24 months with moderate acute respiratory failure (defined by SpO₂/FiO₂ < 264 and signs of respiratory distress). Twelve infants completed the full crossover protocol.
• Intervention: Each infant underwent three 30-minute NIV phases in randomized order, with constant ventilator settings (Pressure Support, PEEP, FiO₂, rise time) and sedation (dexmedetomidine). The three configurations were:
  1. DLC-PICU: A dual-limb compressed-gas pediatric ICU ventilator (Babylog VN500, Dräger).
  2. SLC-Turbine: A turbine-driven ventilator (Astral 150, ResMed) with a single-limb circuit and intentional leak (Whisper valve).
  3. DLC-Turbine: The same turbine-driven ventilator configured with a dual-limb circuit and an active expiratory valve (no intentional leak).
• Measurements:
  • Primary Endpoint: Asynchrony Index (AI%), calculated as the ratio of asynchronous events to total respiratory cycles.
  • Secondary Endpoints: Specific asynchrony events (ineffective efforts, auto-triggering, double triggering, premature/late cycling).
  • Physiological Variables: Esophageal pressure (Pes) was measured via a balloon catheter to assess respiratory muscle effort (ΔPes and Pressure-Time Product, PTPes), breathing patterns, and neural timing.
• Statistical Analysis: Data were analyzed using the Friedman test for repeated measures and Wilcoxon signed-rank tests for pairwise comparisons. Effect sizes were reported as Hodges-Lehmann median differences.

Key Results
The study demonstrated significant differences in patient–ventilator synchrony across the three configurations, though respiratory muscle effort remained comparable.

1. Asynchrony Index (AI):

   • SLC-Turbine achieved the lowest AI (median 7.5% [0–13.7]).
   • DLC-Turbine showed an intermediate AI (16.8% [7.6–25]).
   • DLC-PICU exhibited the highest AI (34.2% [17.4–67]).
   • The difference between DLC-PICU and SLC-Turbine was highly significant (Primary endpoint: +33.0% difference; p=0.0001). The difference between DLC-PICU and DLC-Turbine was also significant (Secondary endpoint: +24.2% difference; p=0.0001).

2. Specific Asynchrony Events:

   • Auto-triggering: Significantly higher with DLC-PICU (4 events/min) compared to SLC-Turbine (0 events/min).
   • Late Cycling: Significantly more frequent with DLC-PICU compared to SLC-Turbine.
   • Ineffective Efforts: Did not differ significantly across any of the three configurations (all near zero).

3. Respiratory Effort and Mechanics:

   • Despite the marked reduction in asynchrony with turbine-driven ventilators, respiratory muscle effort was similar across all three configurations.
   • Esophageal pressure swing (ΔPes) and Pressure-Time Product (PTPes) showed no statistically significant differences between SLC-Turbine and DLC-PICU (p=0.540 and p=0.898, respectively).
   • All ventilator configurations significantly reduced respiratory effort compared to unassisted spontaneous breathing.

4. Breathing Patterns:

   • Patient respiratory rates decreased with all ventilator support compared to spontaneous breathing.
   • Ventilator respiratory rates were significantly higher with DLC-PICU compared to both turbine configurations.

Key Contributions

• Physiological Evidence in Infants: This is the first randomized physiological study to directly compare turbine-driven and compressed-gas ventilators in infants, extending previous adult findings to the pediatric population.
• Disentangling Technology and Circuit: By comparing the DLC-Turbine and SLC-Turbine configurations, the study isolates the effect of circuit architecture. The results suggest that the intentional-leak single-limb configuration contributes independently to superior synchrony, likely due to integrated leak-management strategies.
• Dissociation of Synchrony and Effort: The study provides robust evidence that improved patient–ventilator synchrony (lower AI) does not necessarily equate to reduced inspiratory muscle effort. The Asynchrony Index should not be used as a surrogate for work of breathing.
• Phenotype of Asynchrony: The study highlights that different ventilator platforms generate distinct asynchrony phenotypes. The compressed-gas ventilator was characterized by high auto-triggering and late cycling, whereas ineffective efforts were rare across all systems.

Significance and Claims
The authors claim that turbine-driven ventilators, particularly those with intentional-leak single-limb circuits, substantially improve patient–ventilator interaction compared to conventional dual-limb compressed-gas ICU ventilators in infants.

• Clinical Implications: The findings suggest that optimizing NIV in infants requires attention to both ventilator technology and circuit design. The single-limb turbine configuration offers the greatest reduction in asynchrony.
• Practical Utility: Beyond physiological advantages, the single-limb turbine-driven configuration offers practical benefits such as portability and independence from compressed-gas sources. This makes it particularly attractive for emergency departments, interhospital transport, and resource-limited settings.
• Limitations and Future Directions: The authors modestly note that this was a physiological study with a small sample size and short observation period; it was not designed to evaluate clinical outcomes such as NIV failure rates, intubation, or mortality. They state that whether these physiological advantages translate into improved clinical outcomes requires investigation in adequately powered clinical trials.
• Interpretation of AI: The paper emphasizes that the Asynchrony Index must be interpreted alongside respiratory effort and the specific type of asynchrony events, as high asynchrony burdens in adults (often linked to poor outcomes) may manifest differently in infants (e.g., auto-triggering vs. ineffective efforts).