Closed-loop non-invasive ventilation in critical care: current evidence and clinical perspectives

Abstract

Non-invasive ventilation (NIV) is a cornerstone of acute and perioperative respiratory support, but its effectiveness is constrained by patient selection, interface leaks, patient-ventilator asynchrony, excessive inspiratory effort, and delayed recognition of treatment failure. Closed-loop and adaptive approaches may address some of these limitations by allowing ventilators or decision-support systems to adjust support according to physiological feedback. This mini review synthesizes current evidence on closed-loop concepts relevant to NIV in critical care, including leak-compensation algorithms, automated triggering and cycling, waveform-guided synchronization, proportional and neurally adjusted ventilatory assistance, volume-assured pressure support, monitoring of respiratory drive and effort, and emerging artificial intelligence approaches. The available literature suggests that most clinically available NIV technologies should be regarded as partial closed-loop systems rather than fully autonomous physiological controllers. Evidence is strongest for improved patient-ventilator interaction through leak-adapted algorithms, dedicated NIV ventilators, NAVA, and automated waveform-guided cycling, whereas direct evidence that closed-loop NIV improves patient-centred outcomes in critically ill adults remains limited. Physiological monitoring, particularly early assessment of inspiratory effort, may provide key input signals for future closed-loop NIV strategies designed to personalize support and avoid delayed intubation. Artificial intelligence, ontologies, digital twins, and reinforcement learning may accelerate development, but their current role is mainly preclinical or conceptual. Future trials should test clinically supervised, safety-constrained systems that integrate ventilator waveforms, gas exchange, leaks, comfort, drive, and failure prediction.